Machine learning (ML) accelerators increasingly demand architectural flexibility to support diverse workloads. While reconfigurable designs like Self-Adaptive Reconfigurable Arrays (SARA) improve adaptability, they suffer from physical design bottlenecks-particularly in memory-intensive settings where dense memory integration inflates area, wirelength, and energy consumption. To address these challenges, we propose a scalable multi-tier memory-on-logic (MoL) integration framework that co-optimizes macro placement and standard-cell layout for flexible ML accelerators. Our methodology combines architectural insights with subgradient-based convex optimization for global placement, followed by simulated annealing for overlap resolution and multi-tier partitioning. The proposed flow significantly improves area efficiency, routing congestion, and timing closure. Experimental results on ML workloads demonstrate up to 3.4 × runtime speedup, 4.4 × energy-delay product (EDP) improvement, and $2.3 \times$ chip area reduction compared to conventional two-tier MoL baselines, enabling practical deployment of high-performance, adaptive 3D ML accelerators.

State-of-the-art (SOTA) 3D physical design (PD) flows extend commercial 2D place-and-route (P&R) tools to enable signoff-quality 3D IC implementation through double metal stacking and inter-die metal layer sharing. While metal layer sharing introduces additional routing resources, the substantially higher manufacturing cost of face-to-face (F2F) inter-die vias compared to intra-die vias necessitates 3D-aware routing strategies to manage routability-cost trade-offs. To address this, we propose differentiable routing guidance for 3D ICs (DRG-3D), a GPUaccelerated differentiable optimization framework that provides routing guidance for 3D ICs. DRG-3D formulates a fully differentiable objective that simultaneously optimizes key 3D design metrics: routing congestion, wirelength, via cost, and F 2 F -via cost, which enables efficient and scalable gradient-based optimization over large-scale netlists. Experimental results show that DRG-3D outperforms the SOTA Pin-3D flow, achieving up to 8.37% reduction in routing overflow, 23.99% reduction in total negative slack (TNS), and 18.05% reduction in post-route timing violations.

Backside metal integration effectively mitigates IR-drop challenges, which become increasingly difficult to manage as CMOS technology scales down. However, compared to frontside-only chip designs, implementing backside metal incurs additional costs due to processes such as wafer thinning and nTSV formulation. To maximize the advantages of backside integration, the industry is now exploring functional backside technology. This article introduces a novel BS-PDN-last flow that is critical for the multifunctional utilization of backside metal layers. The BS-PDN-last flow defers Power/Ground (P/G) routing to a later stage of the physical design process, in contrast to the conventional approach where it is performed early in the flow. As a result, BS-PDN-last overcomes the voltage drop and performance trade-offs inherent in conventional flows, enabling substantial performance gains while maintaining low IR-drop levels. Experimental results show that the BS-PDN-last flow achieves a 90% reduction in Total Negative Slack and a 12% performance gain with Backside Clock Delivery Network.

Recent breakthroughs in heterogeneous integration (HI) using 2.5D and 3D ICs have been key to advances in the semiconductor industry. However, heterogeneous integration has also led to several sources of distrust due to the use of thirdparty IP, testing, and fabrication facilities in the design and manufacturing process. Recent work on 2.5D IC security has focused on attacks that can be mounted through rogue chiplets integrated in the design. Thus, existing solutions implement interchiplet communication protocols that prevent unauthorized data modification and interruption in a 2.5D system. However, none of the existing solutions offer inherent security against IP theft. We develop a comprehensive threat model for 2.5D systems indicating that such systems remain vulnerable to IP theft. We present a method that prevents IP theft by obfuscating the connectivity of chiplets on the interposer using reconfigurable interconnection networks. We further achieve dynamic obfuscation of chiplet interconnects while demonstrating resilience against removal, data-snooping, test-based data leakage attacks. We present an approach to implement reconfigurable chiplet interconnection networks through both centralized and distributed scrambler designs, evaluating the security and implementation benefits of both the architectures. We present a comprehensive methodology to design, implement, and integrate interconnect scramblers in a large 2.5D HI system. We also evaluate the power, performance, and area overhead for both distributed and centralized scramblers.

Traditional Design Space Exploration (DSE) methods in Physical Design (PD), such as Bayesian Optimization (BO) and Ant Colony Optimization (ACO), as well as state-of-the-art commercial tools like Synopsys DSO.ai, typically treat the design flow as a black box, lacking insight into the underlying designs. This hinders their ability to generalize across unseen designs. In this article, we introduce FastTuner, an innovative Reinforcement Learning (RL) agent that leverages Graph Neural Networks (GNNs) and Transformers to understand the underlying designs and enable rapid DSE on unseen designs across various PD stages. Our approach incorporates an attention-based framework for autoregressive and conditional parameter tuning and introduces a power, performance and area (PPA) estimator to predict end-of-flow PPA metrics, significantly accelerating RL reward computation. Extensive evaluations on seven industrial designs using the TSMC 28nm technology node demonstrate that FastTuner significantly outperforms existing state-of-the-art DSE techniques in both optimization quality and runtime, achieving improvements of up to 79.38% in Total Negative Slack (TNS), 12.22% in total power, and more than 50x reduction in runtime.

Physical design tools have complex workflows with many different ways of optimizing power, performance, and area (PPA) out of a large number of options and hyperparameters in different engines and functionalities. Black-box optimization techniques are widely adopted to automate quality-of-result (QoR) exploration. Such exploration often proves impractical in real-world customer environments due to high computational demands, lengthy exploration cycles, and the need for large parallel jobs. To reduce the exploration space for viable compute resource requirements, we propose a novel design methodology to enable transferable learning by incorporating design insights crafted on top of physical design experts’ experience and streamlining QoR exploration as a sequence generation task for best recipe selection. We apply language model-inspired alignment techniques to learn the ranking of different recipe sets, enabling our model to generalize beyond known-good manually tuned expert design recipes. Extensive evaluations demonstrate our method’s superior QoRs and runtime performance on unseen industrial designs and rigorous benchmarks.

State-of-the-art 3D IC flows fail to consider 3D congestion during earlier stages, leading to excessive use of end-of-flow ECO resources for routability correction that severely degrades full-chip Power, Performance, and Area metrics. We present DCO-3D, a Machine Learning-based routability-aware 3D PD flow that performs early post-route congestion prediction using Siamese Networks and resolves the predicted hotspots using a fully differentiable 3D cell spreading with Graph Neural Network. On 6 industrial designs in a commercial 3nm node, DCO-3D improves Pin-3D, the known best Pin-3D flow, by up to 47.2% in overflow, 86.2% in TNS and 5.1% in power at signoff.

Native 3D Integrated Circuit (3D IC) design offers enhanced performance and density but faces challenges in signal routing due to limited true 3D EDA tool support. Pseudo-3D flows bridge this gap but lack cross-tier optimization, critical for both mixed-node and homogeneous designs. Metal Layer Sharing (MLS) addresses this by enabling cross-tier routing co-optimization but risks timing degradation if not applied strategically. Additionally, MLS creates open connections in hybrid-bonded 3D ICs, making chips untestable. We propose GNN-MLS, a Graph Neural Network-based framework for precise MLS net selection, combined with a tailored DFT solution for robust testability. Experiments show GNN-MLS reduces timing violations by 79% and improves WNS and TNS by 81% and 94%, and moves designs closer to true 3D ICs.

As the scaling of semiconductor devices nears its limits, utilizing the back-side space of silicon has emerged as a new trend for future integrated circuits. With intense interest, several works have hacked existing backend tools to explore the potential of synthesizing double-side clock trees via nano Through-Silicon-Vias (nTSVs). However, these works lack a systematic perspective on design resource allocation and multi-objective optimization. We propose a systematic approach to design clock trees with double-side metal layers, including hierarchical clock routing, concurrent buffers and nTSVs insertion, and skew refinement. Compared with the state-of-the-art (SOTA) methods, the widely-used open-source tool, our algorithm outperforms them in latency, skew, wirelength, and the number of buffers and nTSVs.

The increasing demand to maximize PPA gains with backside metal layers has driven their function beyond power delivery. This study introduces a novel BS-PDN-last flow, crucial for leveraging multifunctional backside, by deferring power routing to the post-signal routing stage. This approach addresses the IR-drop and performance trade-offs inherent in conventional PDN-first flows. Experimental results show that the BS-PDN-last flow achieves a $90 \%$ reduction in Total Negative Slack and a $\mathbf{1 2 \%}$ performance gain with BS-CDN. Additionally, our work presents the first comprehensive comparison of FS-PDN, BSPDN, and multifunctional backside designs, evaluated on both physical design and workload metrics leveraging accurate vector-based analysis.

Buffer insertion is a critical netlist optimization technique in Physical Design (PD) that balances trade-offs between Power, Performance, and Area (PPA) metrics. Traditional buffering methods rely heavily on local heuristics, which do not scale and often result in globally sub-optimal solutions. Prior Machine Learning (ML) techniques such as BufFormer attempted to alleviate this limitation but remain prohibitively time-consuming (and sub-optimal) due to their incremental nature. In this paper, we introduce BUFFALO, a generative buffer insertion framework that, for the first time in PD, formulates buffer tree generation as a sequence-to-sequence task solved by Large Language Models (LLMs). Particularly, given a design, BUFFALO performs single-shot generation of buffer trees for all fanout-violating nets and INSTA-selected timing critical nets. Furthermore, Group Relative Policy Optimization (GRPO), a Reinforcement Learning (RL) technique, is employed to refine predicted solutions in a PPA-configurable manner. Experimental results on 9 full-chip designs in a 7nm node demonstrate that BUFFALO outperforms an industry-leading commercial PD tool by 71% in Total Negative Slack (TNS), 67.69% in Worst Negative Slack (WNS), and 83x in runtime without incurring additional power consumption.

In the post-Moore era, 3D ICs offer superior timing performance and lower power compared to traditional 2D designs. However, current state-of-the-art (SOTA) flows suffer from excessive snaking along critical paths and require many bond pads. This not only complicates production, but also degrades timing, as each inter-die connection demands a bond pad and 2× BEOL wires. The challenge arises from relying on bin-based FM partitioning, which overlooks inter-bin connections and scales poorly. We introduce Snake-3D, a snaking and timing-aware gradient-based partitioner. It eliminates excessive snaking and concurrently reduces bond pad usage during the partitioning stage, the earliest point these optimizations can be applied. Leveraging GPU acceleration and PyTorch, Snake-3D matches the cutsize quality of advanced gradient methods while running faster on modest hardware. Its integrated snaking- and timing-aware cost, combined with local density control, significantly reduces critical path snaking and achieves superior final PPA compared to GNN-based methods and the default method in SOTA 3D IC flows. On six benchmarks at 7 nm and 28 nm technology nodes, Snake-3D outperforms the default method in SOTA 3D IC flows. It reduces final bond pad by 29%, lowers average and maximum snaking by 34% and 20%, and improves WNS and TNS by 17% and 42%. Compared to GNN-based methods, Snake-3D achieves a 47% reduction in bond pads, reduces snaking by 39% (average) and 30% (maximum), and improves WNS and TNS by 48% and 62%. Snake-3D runs 7.4× faster than GNN-based and similar speed to the default method in SOTA 3D IC flows.

Gate-level 3D ICs have demonstrated substantial power and performance improvements over 2D ICs. However, at advanced technology nodes, achieving sufficient hybrid bond density within a reduced chip footprint requires a 200nm bond pitch—beyond the limits of current manufacturing capabilities. To address this issue, we focus on block-level 3D IC design which has fewer top-level connections. Previous methodologies for block-level 3D IC design suffer from several limitations: (1) reliance on slow simulated annealing-based floorplanning, (2) suboptimal 3D design flows lacking post-route optimization, (3) a lack of attention to the critical initial step of soft block design, and (4) excessive hybrid bond usage that necessitates a 500nm pitch. To overcome these challenges, we propose a comprehensive methodology that includes: (1) a fast, gradient-based analytical solver, (2) Fence-3D flow, delivering up to 15% improvement in power-delay product, (3) ML-based congestion-aware soft block sizing, delivering up to 6.6% improvement, and (4) a partial-MLS method that selectively applies Metal Layer Sharing, reducing hybrid bond count by up to 71%. Collectively, these techniques enable the use of a 1um hybrid bond pitch in 3nm block-level 3D ICs.

Spiking Neural Networks (SNNs) provide a brain-inspired and event-driven mechanism that is believed to be critical to unlock energy-efficient deep learning. On the other hand, mixture-of-experts (MoE) models mirror the parallel distributed processing of the nervous system, and expand model capacity without scaling up the number of computational operations. However, there is currently a lack of hardware support for highly parallel distributed processing in spiking based MoE models. This paper introduces the first 3D hardware architecture and design methodology for Mixture-of-Experts and Multi-Head Attention spiking transformers. By leveraging 3D integration with memory-on-logic and logic-on-logic stacking and exploring energy-efficient dynamic head pruning, we explore such brain-inspired accelerators with spatially stackable circuitry, demonstrating significant improvements of energy efficiency and latency over conventional 2D CMOS integration.

3D Integrated Circuits (ICs) significantly enhance chip performance but require substantial engineering due to their expanded design space. To tackle this, we introduce the MIX-3D framework, an advanced optimizer utilizing Variational Autoencoders for robust extrapolation, identifying energy-efficient and thermal-aware design configurations for mixed-node, mixed-area 3D ICs. It enables architecture-circuit co-design for F2F 2-tier Logic-on-Memory 3D ICs, providing real-time predictions of both back-end and front-end metrics. Additionally, transfer learning reduces dataset construction time by 64%, a common challenge in supervised learning. Experimental results demonstrate that MIX-3D delivers 12% improvements in energy efficiency and 62% less power compared to equal-area 3D ICs. Furthermore, our thermal-aware design reduces chip temperature by 38%.

Back-end-of-line (BEOL) integrated voltage regulators (IVRs) significantly enhance power integrity by reducing on-chip and off-chip dynamic IR-drop compared to traditional off-chip IVR solutions. As semiconductor technologies scale down, power delivery networks increasingly suffer from parasitic limitations, impacting efficiency and performance. While backside power delivery networks (BS-PDNs) offer improvements by mitigating frontside parasitics and power density challenges, conventional architectures remain limited by package-level parasitics in 2.5D interposer systems. To address these constraints, we introduce optimized BEOL IVR structures suitable for BEOL integration, utilizing amorphous tungsten-doped indium oxide (IWO) transistors to enable monolithic 3D integration without additional footprint overhead. Our results demonstrate that BEOL IVRs achieve up to 61% and 54% reductions in on-chip dynamic IR-drop for frontside and backside configurations, respectively, with backside BEOL IVRs providing an additional 18% improvement. Furthermore, BEOL IVRs substantially decrease off-chip dynamic IR-drop by up to 71%, presenting a scalable and technology-node-independent solution for advanced IC designs.

Hybrid bonding is emerging as a key technology for 3D integration, offering finer bonding pitches that address the high interconnect density requirements of modern VLSI applications. In advanced node technologies, where metal pitches are significantly smaller than bonding pitches, 3D net assignment becomes critical for achieving optimal design performance. Existing approaches primarily focus on either ensuring the legality of the assignment or optimizing the flexibility of 3D net locations for timing purposes in isolation. This limitation restricts the performance improvements of 3D designs over traditional 2D counterparts. To overcome these challenges, we introduce AnchorGrid, a novel 3D net assignment framework designed to concurrently assign 3D nets to legal locations while supporting their movement to enhance timing optimization. By modeling 3D nets as pairs of specialized ''anchor'' cells, accompanied by relative placement constraints, precise movement and alignment are achieved during the pre-route optimization phase, before final placement onto grid-based locations. Experimental results on advanced node commercial designs demonstrate that AnchorGrid achieves up to a 24.35% improvement in power, performance, and area (PPA) metrics, while reducing design rule check (DRC) violations by 90%, outperforming state-of-the-art methods.

Glass interposers have become a compelling option for 2.5-D heterogeneous integration compared to silicon. It allows 3-D stacking configuration between the embedded dies and the conventional flip-chip dies mounted directly on top at low cost. Furthermore, the interconnect pitch and through-glass-via (TGV) diameter in glass are becoming comparable to their counterparts in silicon. In this study, we investigate the power, performance, area (PPA), signal integrity (SI) and power integrity (PI) advantages of 3-D stacking afforded by glass interposers over silicon interposers. Our research employs a chiplet/package co-design approach, progressing from an register-transfer-level description of RISC-V chiplets to final graphic data system (GDS) layouts, utilizing TSMC 28 nm for chiplets and Georgia Tech’s 3-D glass packaging for the interposer. Compared to silicon, glass interposers offer a$2.6\times $reduction in area, a$21\times $reduction in wire length, a 17.72% reduction in full-chip power consumption, a 64.7% increase in SI and a$10\times $improvement in PI, with a 35% increase in thermal. Furthermore, we provide a detailed comparative analysis with 3-D Silicon technologies. It not only highlights the competitive advantages of glass interposers, but also provides critical insights into each design’s potential limitations and optimization opportunities.

This article presents 3DNN-Xplorer, the first machine learning (ML)-based framework for predicting the performance of heterogeneous 3-D deep neural network (DNN) accelerators. Our ML framework facilitates the design space exploration (DSE) of heterogeneous 3-D accelerators with a two-tier compute-on-memory (CoM) configuration, considering 3-D physical design factors. Our design space encompasses four distinct heterogeneous 3-D integration styles, combining 28- and 16-nm technology nodes for both compute and memory tiers. Using extrapolation techniques with ML models trained on 10-to-256 processing element (PE) accelerator configurations, we estimate the performance of systems featuring 75–16384 PEs, achieving a maximum absolute error of 13.9% (the number of PEs is not continuous and varies based on the accelerator architecture). To ensure balanced tier areas in the design, our framework assumes the same number of PEs or on-chip memory capacity across the four integration styles, accounting for area imbalance resulting from different technology nodes. Our analysis reveals that the heterogeneous 3-D style with 28-nm compute and 16-nm memory is energy-efficient and offers notable energy savings of up to 50% and an 8.8% reduction in runtime compared to other 3-D integration styles with the same number of PEs. Similarly, the heterogeneous 3-D style with 16-nm compute and 28-nm memory is area-efficient and shows up to 8.3% runtime reduction compared to other 3-D styles with the same on-chip memory capacity.

The back-side metal layers exhibit lower parasitics compared to the front-side layers in advanced technologies, making them suitable for clock-net distribution. In this study, we explore the advantages of using back-side metal layers for clock routing, which is shared with a power delivery network. Our Graph Neural Network (GNN) based framework, effectively distributes the clock-tree between the front and back sides. We address the back-side clock nets' creation by incorporating back-side buffers. Our results demonstrate better clock and full-chip metrics represented by an increase of up to 13% in the effective frequency with equivalent power consumption, using 3 nm technology.

While various studies have shown effective parameter optimizations for specific designs, there is limited exploration of parameter optimization within the domain of 3D Integrated Circuits. We present the first comprehensive study, both qualitatively and quantitatively, comparing five state-of-the-art (SOTA) techniques for parameter optimization applied to 3D ICs. Additionally, we introduce an end-to-end machine learning-based framework, encompassing important parameter selection through optimization, all without human intervention. Extensive studies across six industrial designs under the TSMC 28nm technology node reveal that our proposed framework outperforms SOTA techniques in three different optimization objectives in both optimization quality and runtime.

This study presents a novel approach to address the challenges of scaling limitations and high read/write latencies inherent in conventional 2D SRAM designs through the development of a 15nm FinFET-based standalone 3D SRAM Array. Utilizing Monolithic Intertier Vias, our two-tiered implementation effectively mitigates these limitations along both X and Y axes. We introduce two distinct design methodologies for 3D SRAM arrays - Wordline and Bitline folding. Through post-layout simulations conducted on arrays with capacities of 2kB (256WLx64BL) and 8kB (512WLx128BL), our 3D designs exhibit superior performance metrics. Notably, the 3D Wordline-Folded design achieves a remarkable 57.54% average reduction in footprint compared to the 2D baseline across both array sizes. Furthermore, the 3D Bitline-Folded configuration demonstrates consistent superiority in speed, with an average read latency improvement of 17.16% and a notable 54.2% enhancement in write latency. Conversely, the 3D Wordline-Folded array emerges as the most energy-efficient option, boasting an average reduction of 15.3% in read energy and over 21% in write energy compared to the 2D baseline.

2.5D integration is gaining popularity primarily due to its ability to facilitate intellectual property (IP) reuse. Unlike conventional 2D and 3D approaches, 2.5D integration requires a more complex design and analysis process and is highly sensitive to changes in design parameters. However, research on the sensitivity of 2.5D design parameters is notably scarce, with most studies still concentrating on 2D and 3D. In this paper, we propose an AI-driven model for predicting sensitivity and an optimization methodology for 2.5D parameters, with a particular focus on bump pitch. Our approach employs advanced machine learning models to accurately predict how variations in bump pitch impact the power, performance, and area of chiplets, as well as the footprint, signal integrity, power integrity, and thermal integrity of the interposer layout. We also utilize Bayesian optimization to identify the optimal bump pitch for specific design objectives. Experimental validation of our model demonstrates high accuracy, with average relative errors of 2.69% for interpolation and 2.7% for extrapolation. Furthermore, optimization results, tailored by adjusting weights for various potential design goals, show an average improvement of 11% in area, wire length, and signal integrity-driven optimization, and 9% in power and thermal integrity-driven optimization.

Spiking neural networks (SNNs) are powerful models of spatiotemporal computation and are well suited for deployment on resource-constrained edge devices and neuromorphic hardware due to their low power consumption. Leveraging attention mechanisms similar to those found in their artificial neural network counterparts, recently emerged spiking transformers have showcased promising performance and efficiency by capitalizing on the binary nature of spiking operations. Recognizing the current lack of dedicated hardware support for spiking transformers, this paper presents the first work on 3D spiking transformer hardware architecture and design methodology. We present an architecture and physical design co-optimization approach tailored specifically for spiking transformers. Through memory-on-logic and logic-on-logic stacking enabled by 3D integration, we demonstrate significant energy and delay improvements compared to conventional 2D CMOS integration.

Heterogeneous 3D integration, blending multiple technology nodes, emerges as a promising strategy for enhancing performance and maintaining low power consumption in next-generation computing systems. This paper presents Hetero-3D, an RTL-to-GDS design flow tailored specifically for heterogeneous 3D ICs. Hetero-3D integrates an area-unbalanced 3D floorplanner with an ML-based power delivery and signal router, working in tandem for rigorous PPA (Power, Performance, and Area) optimization. Using two CPU benchmarks, we showcase a remarkable 15% increase in maximum frequency and a substantial 50% decrease in voltage drop compared to homogeneous 3D baselines. Moreover, Hetero-3D effectively addresses voltage drop issues in the power delivery network while delivering an additional 5% frequency boost. This study emphasizes the EDA solutions that unlock the potential of mixed-node stacking as a crucial enabler for performance scaling in future ICs.

Current state-of-the-art Design Space Exploration (DSE) methods in Physical Design (PD), including Bayesian optimization (BO) and Ant Colony Optimization (ACO), mainly rely on black-boxed rather than parametric (e.g., neural networks) approaches to improve end-of-flow Power, Performance, and Area (PPA) metrics, which often fail to generalize across unseen designs as netlist features are not properly leveraged. To overcome this issue, in this paper, we develop a Reinforcement Learning (RL) agent that leverages Graph Neural Networks (GNNs) and Transformers to perform "fast" DSE on unseen designs by sequentially encoding netlist features across different PD stages. Particularly, an attention-based encoder-decoder framework is devised for "conditional" parameter tuning, and a PPA estimator is introduced to predict end-of-flow PPA metrics for RL reward estimation. Extensive studies across 7 industrial designs under the TSMC 28nm technology node demonstrate that the proposed framework FastTuner, significantly outperforms existing state-of-the-art DSE techniques in both optimization quality and runtime. where we observe improvements up to 79.38% in Total Negative Slack (TNS), 12.22% in total power, and 50x in runtime.

Hierarchical very-large-scale integration (VLSI) flows are an understudied yet critical approach to achieving design closure at giga-scale complexity and gigahertz frequency targets. This paper proposes a novel hierarchical physical design flow enabling the building of high-density and commercial-quality two-tier face-to-face-bonded hierarchical 3D ICs. Complemented with an automated floorplanning solution, the flow allows for system-level physical and architectural exploration of 3D designs. As a result, we significantly reduce the associated manufacturing cost compared to existing 3D implementation flows and, for the first time, achieve cost competitiveness against the 2D reference in large modern designs. Experimental results on complex industrial and open manycore processors demonstrate in two advanced nodes that the proposed flow provides major power, performance, and area/cost (PPAC) improvements of 1.2 -2.2× compared with 2D, where all metrics are improved simultaneously, including up to 20% power savings.

As state-of-the-art 3D IC Place-and-Route flows were designed with older technology nodes and aggressive bonding pitch assumptions, they introduce an unacceptable number of 3D via overlap violations during routing in real-world scenarios. Specifically, when dealing with higher via pitch to wire size ratios using more advanced technology nodes than they were designed for, these flows struggle to comply with width and spacing rules. In this paper, we propose a novel 3D via legalization stage and a subsequent refinement stage during routing to address this issue. Two independent via legalization methods are introduced: a force-based algorithm and a bipartite-matching algorithm with Bayesian optimization. Our two legalization methods, along with the refinement stage, are compatible with various process nodes, bonding technologies, and partitioning styles. By implementing the modified 3D routing with the proposed legalizers, we successfully eliminate all 3D via overlap violations while minimizing the impact on performance, power, or area.

Three Dimensional fabrication and packaging of Integrated Circuits has been proposed as one of the key drivers for More Moore technologies by the IRDS. Such integration is useful to improve the performance and cost effectiveness of the newer generation of chips. Several consumer chips have been using micro-bump-based 3-D package bonding techniques but such integration is only done at a very high level. To fully utilize the benefits of 3-D integration, we propose an effective optimization methodology for 3-D ICs. In this work, we present an all-round physical design methodology to support 3-D IC timing optimization, with features, such as timing driven placement, clock tree synthesis, 3-D timing optimization, and ECO optimization for 3-D ICs.

Recently, GPU-accelerated placers such as DREAMPlace and Xplace have demonstrated their superiority over traditional CPU-reliant placers by achieving orders of magnitude speed up in placement runtime. However, due to their limited focus in placement objectives (e.g., wirelength and density), the placement quality achieved by DREAMPlace or Xplace is not comparable to that of commercial tools. In this article, to bridge the gap between open source and commercial placers, we present a novel placement optimization framework named GAN-Place that employs generative adversarial learning to transfer the placement quality of the industry-leading commercial placer, Synopsys ICC2, to existing open source GPU-accelerated placers (DREAMPlace and Xplace). Without the knowledge of the underlying proprietary algorithms or constraints used by the commercial tools, our framework facilitates transfer learning to directly enhance the open source placers by optimizing the proposed differentiable loss that denotes the “similarity” between DREAMPlace- or Xplace-generated placements and those in commercial databases. Experimental results on seven industrial designs not only show that our GAN-Place immediately improves the Power, Performance, and Area metrics at the placement stage but also demonstrates that these improvements last firmly to the post-route stage, where we observe improvements by up to 8.3% in wirelength, 7.4% in power, and 37.6% in Total Negative Slack on a commercial CPU benchmark.

In this article, we present three commercial-grade 3-D IC designs based on state-of-the-art design technologies, specifically microbumping (3-D die stacking), hybrid bonding (wafer-on-wafer bonding), and monolithic 3-D (M3D) ICs. To highlight tradeoffs present in these three designs, we perform analyses on power, performance, and area (PPA) and the clock tree. We also model the tier-to-tier interconnection in each 3-D IC methodology and analyze signal integrity (SI) to assess the reliability of each design. From our experiments using the OpenPiton benchmark, the hybrid bonding design shows the best timing improvement of 81.4% when compared to its 2-D counterpart, while microbumping shows the best reliability among 3-D IC designs. Moreover, we expand our study to the commercial processor architecture, which is Arm Cortex-A53, with the new set of 3-D integration options. In addition, we show the microbump assignment methodology to handle a large number of 3-D interconnections in the microbumping 3-D design. We also perform SI on the new set of 3-D intertier/interdie connections to discuss the reliability based on their physical dimensions. With a new benchmark design, the hybrid-bonding 3-D shows the best energy–delay-product (EDP) improvement, which is 25.8% compared to 2-D, and the largest eye-opening among 3-D integration options.

Monolithic 3-D (M3D) integrated circuits (ICs) introduce a new aspect to the physical design problem by stacking dies vertically. We introduce a novel heterogeneous design for M3D ICs, which utilizes different technology processes for each die. We also propose enhancements to the design flow and improved partitioning methods to support our suggested heterogeneous 3-D IC design. We perform 3-D partitioning with low-cost, low-power, and low-performance cells on the bottom die, and vice versa on the top die. This arrangement allows us to create a high-performance 3-D IC that is more cost-effective and power-efficient than both the 3-D and 2-D homogeneous implementations. We demonstrate these advantages using four different netlists, showing up to a 23% improvement in performance per cost (PPC), and a 16% improvement in power delay product (PDP) when comparing our heterogeneous M3D design to the best 2-D designs.

Concurrent Clock and Data (CCD) optimization is a well-adopted approach in modern commercial tools that resolves timing violations using a mixture of clock skewing and delay fixing strategies. However, existing CCD algorithms are flawed. Particularly, they fail to prioritize violating endpoints for different optimization strategies correctly, leading to flow-wise globally sub-optimal results. In this paper, we overcome this issue by presenting RL-CCD, a Reinforcement Learning (RL) agent that selects endpoints for useful skew prioritization using the proposed EP-GNN, an endpoint-oriented Graph Neural Network (GNN) model, and a Transformer-based self-supervised attention mechanism. Experimental results on 19 industrial designs in 5 − 12nm technologies demonstrate that RL-CCD achieves up to 64% Total Negative Slack (TNS) reduction and 66.5% number of violating endpoints (NVE) improvement over the native implementation of a commercial tool.

Glass interposers enable 3D stacking between the chiplets embedded into the substrate and the ones stacked directly on top, which is not possible in silicon. In this work, we demonstrate the benefits of such stacking in glass interposers over silicon in terms of key system-level metrics including area, wirelength, signal, power, and thermal integrity. We achieve this goal with GDS layouts of both chiplets and interposers and sign-off simulations. Our experiments show that glass offers 2.6X area, 21X wirelength, 17.72% full-chip power, 64.7% signal integrity, and 10X power integrity improvement over silicon at the cost of 15% increase in temperature.

In this paper, we provide an overview of design automation tools and methodology for Monolithic 3D ICs, focusing on the accomplishments in recent years and the gaps that remain to be filled. Monolithic 3D integration is an emerging technology with high 3D interconnect density and performance benefits, but it proposes new challenges for computer-aided design tools. In this paper, we first revisit the current status of design automation tools for Monolithic 3D and highlight the recent developments in tier partitioning, 3D placement and routing, inter-tier via controls, and power and thermal integrity analysis. Then, we discuss the gaps to be met for next-generation system-level heterogeneous Monolithic 3D IC design. Finally, we present our vision for the future of design automation developments for Monolithic 3D ICs.

This paper focuses on iso-on-chip-memory-capacity and iso-footprint Energy-Delay-Product (EDP) benefits of ultra-dense 3D, e.g., monolithic 3D (M3D), computing systems vs. corresponding 2D designs. Simply folding existing 2D designs into corresponding M3D physical designs yields limited EDP benefits$(\sim 1.4\times)$. New M3D architectural design points that exploit M3D physical design are crucial for large M3D EDP benefits. We perform comprehensive architectural exploration and detailed M3D physical design using foundry M3D process design kit and standard cell library for front-end-of-line (FEOL) Si CMOS logic, on-chip back-end-of-line (BEOL) memory, and a single layer of on-chip BEOL FETs. We find new M3D AI/ML accelerator architectural design points that have iso-footprint, iso-on-chip-memory-capacity EDP benefits ranging from$5.3\times$to$11.5\times$vs. corresponding 2D designs (containing only FEOL Si CMOS and on-chip BEOL memory). We also present an analytical framework to derive architectural insights into these benefits, showing that our principles extend to many architectural design points across various device technologies.

Recent breakthroughs in heterogeneous integration (HI) technologies using 2.5D and 3D ICs have been key to advances in the semiconductor industry. However, heterogeneous integration has also led to several sources of distrust due to the use of third-party IP, testing, and fabrication facilities in the design and manufacturing process. Recent work on 2.5D IC security has only focused on attacks that can be mounted through rogue chiplets integrated in the design. Thus, existing solutions implement inter-chip let communication protocols that prevent unauthorized data modification and interruption in a 2.5D system. However, none of the existing solutions offer inherent security against IP theft. We develop a comprehensive threat model for 2.5D systems indicating that such systems remain vulnerable to IP theft. We present a method that prevents IP theft by obfuscating the connectivity of chiplets on the interposer using reconfigurable interconnection networks. We also evaluate the PPA impact and security offered by our proposed scheme.

This paper presents 3DNN-Xplorer, the first machine learning (ML)-based framework for predicting the performance of heterogeneous 3D DNN accelerators. Our ML frame-work facilitates the design space exploration of heterogeneous 3D accelerators with a 2-tier compute-on-memory configuration, considering 3D physical design factors. Our design space encompasses four distinct heterogeneous 3D integration styles, combining 28nm and 16nm technology nodes for both compute and memory tiers. Using extrapolation techniques with ML models trained on 16, 32, and 64 PE accelerator configurations, we estimate the performance of systems featuring 128, 256, 512, 1024, and 2048 PEs, achieving a maximum absolute error of 12.7%. To ensure balanced tier areas in the design, our framework assumes the same number of PEs or on-chip memory capacity across the four integration styles, accounting for area imbalance resulting from different technology nodes. Our analysis reveals that the heterogeneous 3D style with 28nm compute and 16nm memory is energy-efficient and offers notable energy savings of up to 50% and an 8.8% reduction in runtime compared to other 3D integration styles with the same number of PEs. Similarly, the heterogeneous 3D style with 16nm compute and 28nm memory is area-efficient and shows up to 8.3% runtime reduction compared to other 3D styles with the same on-chip memory capacity.

Low latency inference has many applications in edge machine learning. In this paper, we present a run-time configurable convolutional neural network (CNN) inference ASIC design for low-latency edge machine learning. By implementing a 5-stage pipelined CNN inference model in a 3D ASIC technology, we demonstrate that the model distributed on two dies utilizing face-to-face (F2F) 3D integration achieves superior performance. Our experimental results show that the design based on 3D integration achieves 43% better energy-delay product when compared to the traditional 2D technology.

The standard cells are becoming increasingly smaller due to aggressive device down-scaling, and power rails take up a sizable portion of the available space. Buried Power Rail (BPR) and Back-Side Power (BSP) have been gaining more attention owing to their capacity to reduce the standard cell height from 6-Track in the traditional Front Side Power Rail (FS-PR) to 5-Track and 4-Track, respectively. In this paper, we provide a comprehensive comparison of power rail topologies at the device, standard cell, and full chip design level in terms of Power, Performance and Area (PPA). Our experiments show that nanosheet width scaling for BPR and BSP reduces device gate capacitance by 26% and 40%, respectively, resulting in an improvement of internal power of over 33% and 40%, respectively, at the standard cell level, and total power drop of over 24% and 30%, respectively, at the full chip level. Additionally, the floorplan can be shrunk down by 7% with BPR compared to FSPR, and even further by an additional 17% with BSP. This study also demonstrates the Back-side Power delivery network (BS-PDN) benefits in IR drop for BPR and BSP topologies.

DREAMPlace is a renowned open-source placer that provides GPU-acceleratable infrastructure for placements of Very-Large-Scale-Integration (VLSI) circuits. However, due to its limited focus on wirelength and density, existing placement solutions of DREAMPlace are not applicable to industrial design flows. To improve DREAMPlace towards commercial-quality without knowing the black-boxed algorithms of the tools, in this paper, we present DREAM-GAN, a placement optimization framework that advances DREAMPlace using generative adversarial learning. At each placement iteration, aside from optimizing the wirelength and density objectives of the vanilla DREAMPlace, DREAM-GAN computes and optimizes a differentiable loss that denotes the similarity score between the underlying placement and the tool-generated placements in commercial databases. Experimental results on 5 commercial and OpenCore designs using an industrial design flow implemented by Synopsys ICC2 not only demonstrate that DREAM-GAN significantly improves the vanilla DREAMPlace at the placement stage across each benchmark, but also show that the improvements last firmly to the post-route stage, where we observe improvements by up to 8.3% in wirelength and 7.4% in total power.

State-of-the-art 3D IC Place-and-Route flows were designed with older technology nodes and aggressive bonding pitch assumptions. As a result, these flows fail to honor the width and spacing rules for the 3D vias with realistic pitch values. We propose a critical new 3D via legalization stage during routing to reduce such violations. A force-based solver and bipartite-matching algorithm with Bayesian optimization are presented as viable legalizers and are compatible with various process nodes, bonding technologies, and partitioning types. With the modified 3D routing, we reduce the 3D via violations by more than 10× with zero impact on performance, power, or area.

Critical to achieving power–performance–area goals, a human engineer typically spends a considerable amount of time tuning the multiple settings of a commercial placer. This article proposes a deep reinforcement learning (RL) framework to optimize the placement parameters of a commercial electronic design automation (EDA) tool. We build an autonomous agent that learns to tune parameters without human intervention and domain knowledge, trained solely by RL from self-search. To generalize to unseen netlists, we use a mixture of handcrafted features from graph topology theory and graph embeddings generated using unsupervised graph neural networks. Our RL algorithms are chosen to overcome the sparsity of data and latency of placement runs. As a result, our trained RL agent achieves up to 11% and 2.5% wire length improvements on unseen netlists compared with a human engineer and a state-of-the-art tool auto-tuner in just one placement iteration ($20\times $and$50\times $fewer iterations). In addition, the success of the RL agent is measured using a statistical test with theoretical guarantees and an optimized sample size.

Monolithic 3-D (M3D) integration is a promising technology for achieving high performance and low-power consumption. However, the limitations of current M3D fabrication flows lead to performance degradation of devices in the top tier and unreliable interconnects between tiers. Fault localization at the tier level is therefore necessary to enhance yield learning, For example, tier-level localization can enable targeted diagnosis and process optimization efforts. In this article, we develop a graph neural network-based diagnosis framework to efficiently localize faults to a device tier. The proposed framework can be used to provide rapid feedback to the foundry and help enhance the quality of diagnosis reports generated by commercial tools. Results for four M3D benchmarks, with and without response compaction, show that the proposed solution achieves up to 32.86% improvement in diagnostic resolution with less than 1% loss of accuracy, compared to results from commercial tools. The proposed framework has also been demonstrated to be transferable to perform diagnosis on various design configurations without performance degradation.

3-D integration technology is one of the leading options to advance Moore’s Law beyond conventional scaling. One of the 3-D integration choice is the heterogeneous integration with the benefits of power saving over the homogeneous integration. With the lack of commercial 3-D tools, existing 3-D physical design flows utilize 2-D commercial tools to perform 3-D integrated circuit (3-D IC) physical synthesis. Specifically, these flows build 2-D designs first and then convert them into 3-D designs. However, several works demonstrate that design qualities degrade during this 2-D–3-D transformation and some of the flows do not support heterogeneous integration. In this article, we propose Snap-3D, a constraint-driven placement approach to build commercial-quality 3-D ICs, which supports both homogeneous and heterogeneous 3-D ICs. Our key idea is based on the observation that if the standard cell height is contracted and partitioned into multiple tiers, any commercial 2-D placer can place them onto the row structure and naturally achieve high-quality 3-D placement. This methodology is shown to optimize power, performance, and area (PPA) metrics across different tiers simultaneously and minimize the aforementioned design quality loss. Experimental results on seven industrial designs demonstrate that Snap-3D achieves up to 10.9% wirelength, 9% power, and 25% performance improvements compared with state-of-the-art 3-D design flows.

Modern electronic design automation flows depend on both implementation and signoff tools to perform timing-constrained power optimization through Engineering Change Orders (ECOs), which involve gate sizing and threshold-voltage ( V th )-assignment of standard cells. However, the signoff ECO optimization is highly time-consuming, and the power improvement is hard to predict in advance. Ever since the industrial benchmarks released by the ISPD-2012 gate-sizing contest, active research has been conducted extensively to improve the optimization process. Nonetheless, previous works were mostly based on heuristics or analytical methods whose timing models were oversimplified and lacked of formal validations from commercial signoff tools. In this article, we propose ECO-graph neural networks (GNN), a transferable graph-learning-based framework, which harnesses GNNs to perform commercial-quality signoff power optimization through discrete ( V th -assignment. One of the highlights of our framework is that it generates tool-accurate optimization results instantly on unseen netlists that are not utilized in the training process. Furthermore, we propose a subgraph approximation technique to improve training and inferencing time of the proposed GNN model. We show that design instances with non-overlapping subgraphs can be optimized in parallel so as to improve the inference time of the learning-based model. Finally, we implement a GNN-based explanation method to interpret the optimization results achieved by our framework. Experimental results on 14 industrial designs, including a RISC-V-based multi-core system and the renowned ISPD-2012 benchmarks, demonstrate that our framework achieves up to 14× runtime improvement with similar signoff power optimization quality compared with Synopsys PrimeTime , an industry-leading signoff tool.

D Place and Route (P&R) flows either involve true-3D placement algorithms or use commercial 2D tools to transform a 2D design into a 3D design. Irrespective of the nature of the placers, several placement parameters in these tools affect the quality of the final 3D designs. Different parameter settings work well with different circuits, and it is impossible to manually tune them for a particular circuit. Automated approaches involving reinforcement learning have been shown to adapt and learn the parameter settings and create trained models. However, their effectiveness depends on the input dataset quality. Using a set of 10 netlists and 10–21 handpicked placement parameters in P&R flows involving pseudo-3D or true-3D placement, the dataset quality is analyzed. The datasets are the design metrics obtained through different P&R stages, such as placement optimization, clock tree synthesis, or 3D partitioning and global routing. The training runtime and the quality of the final design metrics are compared. On a pseudo-3D flow, the training takes around 126–290 hours, whereas, on a true-3D placer-based flow, it takes around 305–410 hours. It is observed that the datasets obtained from different stages lead to drastically different final design results. With the RL-based training processes, the quality of results in 3D designs improves by up to 23.7% compared to their corresponding untrained P&R flows.

In high-performance three-dimensional Integrated Circuits (3D ICs), clock networks consume a large portion of the full-chip power. However, no previous 3D IC work has ever optimized 3D clock networks for both power and performance simultaneously, which results in sub-optimal 3D designs. To overcome this issue, in this article, we propose a GNN-based flip-flop clustering algorithm that merges single-bit flip-flops into multi-bit flip-flops in an unsupervised manner, which jointly optimizes the power and performance metrics of clock networks. Moreover, we integrate our algorithm into the state-of-the-art 3D physical design flow and verify the integration, which leads to a better 3D full-chip design. Experimental results on eight industrial benchmarks demonstrate that the algorithm achieves improvements up to 18% in total power and 8.2% in performance over the state-of-the-art 3D flow.

Nanoscale interlayer vias (ILVs) in monolithic 3-D (M3D) ICs have enabled high-density vertical integration of logic and memory tiers. However, the sequential assembly of M3D tiers via wafer bonding is prone to variability in the immature fabrication process and manufacturing defects. The yield degradation due to ILV faults can be mitigated via dedicated test and diagnosis of ILVs using built-in self-test (BIST). Prior work has carried out fault localization for a regular 1-D placement of ILVs in the M3D layout where shorts are assumed to arise only between unidirectional ILVs. However, to minimize wirelength in M3D routing, ILVs may be irregularly placed by a place-and-route tool, and shorts can also occur between an up-going ILV and a down-going ILV. To test and localize faults in realistic ILV layouts, we present a new BIST framework that is optimized for test time and PPA overhead. We also present a graph-theoretic approach for representing potential fault sites in the ILVs and carry out inductive fault analysis to drop noncritical sites. We describe a procedure for optimally assigning ILVs to the BIST pins and determining the BIST configuration for test-cost minimization. Evaluation results for M3D benchmarks demonstrate the effectiveness of the proposed framework.

This work identifies the architectural and design scaling limits of 2-D flexible interconnect deep neural network (DNN) accelerators and addresses them with 3-D ICs. We demonstrate how scaling up a baseline 2-D accelerator in the$X/Y$dimension fails and how vertical stacking effectively overcomes the failure. We designed multitier accelerators that are$1.67\times $faster than the 2-D design. Using our 3-D architecture and circuit codesign methodology, we improve throughput, energy efficiency, and area efficiency by up to$5\times $,$1.2\times $, and$3.9\times $, respectively, over 2-D counterparts. The IR-drop in our 3-D designs is within 10.7% of VDD, and the temperature variation is within 12 °C.

Three-dimensional integrated circuits promise power, performance, and footprint gains compared to their 2D counter-parts, thanks to drastic reductions in the interconnects' length through their smaller form factor. We can leverage the potential of 3D integration by enhancing MemPool, an open-source many-core design with 256 cores and a shared pool of L1 scratchpad memory connected with a low-latency interconnect. MemPool's baseline 2D design is severely limited by routing congestion and wire propagation delay, making the design ideal for 3D integration. In architectural terms, we increase MemPool's scratchpad memory capacity beyond the sweet spot for 2D designs, improving performance in a common digital signal processing kernel. We propose a 3D MemPool design that leverages a smart partitioning of the memory resources across two layers to balance the size and utilization of the stacked dies. In this paper, we explore the architectural and the technology parameter spaces by analyzing the power, performance, area, and energy efficiency of MemPool instances in 2D and 3D with 1 MiB, 2 MiB, 4 MiB, and 8 MiB of scratchpad memory in a commercial 28 nm technology node. We observe a performance gain of 9.1% when running a matrix multiplication on MemPool-3D with 4 MiB of scratchpad memory compared to the MemPool 2D counterpart. In terms of energy efficiency, we can implement the MemPool-3D instance with 4 MiB of L1 memory on an energy budget 15 % smaller than its 2D counterpart, and 3.7 % smaller than the MemPool-2D instance with a fourth of the L1 scratchpad memory capacity.

As modern Physical Design (PD) algorithms and methodologies evolve into the post-Moore era with the aid of machine learning, Graph Neural Networks (GNNs) are becoming increasingly ubiquitous given that netlists are essentially graphs. Recently, their ability to perform effective graph learning has provided significant insights to understand the underlying dynamics during netlist-to-layout transformations. GNNs follow a message-passing scheme, where the goal is to construct meaningful representations either at the entire graph or node-level by recursively aggregating and transforming the initial features. In the realm of PD, the GNN-learned representations have been leveraged to solve the tasks such as cell clustering, quality-of-result prediction, activity simulation, etc., which often overcome the limitations of traditional PD algorithms. In this work, we first revisit recent advancements that GNNs have made in PD. Second, we discuss how GNNs serve as the backbone of novel PD flows. Finally, we present our thoughts on ongoing and future PD challenges that GNNs can tackle and succeed.

Hierarchical very-large-scale integration (VLSI) flows are an understudied yet critical approach to achieving design closure at giga-scale complexity and gigahertz frequency targets. This paper proposes a novel hierarchical physical design flow enabling the building of high-density and commercial-quality two-tier face-to-face-bonded hierarchical 3D ICs. We significantly reduce the associated manufacturing cost compared to existing 3D implementation flows and, for the first time, achieve cost competitiveness against the 2D reference in large modern designs. Experimental results on complex industrial and open manycore processors demonstrate in two advanced nodes that the proposed flow provides major power, performance, and area/cost (PPAC) improvements of 1.2 to 2.2 × compared with 2D, where all metrics are improved simultaneously, including up to power savings.

Micro-bump and hybrid bonding technologies have enabled 3D ICs and provided remarkable performance gain, but the memory macro partitioning problem also becomes more complicated due to the limited 3D connection density. In this paper, we evaluate and quantify the impacts of various macro partitioning on the performance and temperature in commercial-grade 3D ICs. In addition, we propose a set of partitioning guidelines and a quick constraint-graph-based approach to create floorplans for logic-on-memory 3D ICs. Experimental results show that the optimized macro partitioning can help improve the performance of logic-on-memory 3D ICs by up to 15%, at the cost of 8°C temperature increase. Assuming air cooling, our simulation shows the 3D ICs are thermally sustainable with 97°C maximum temperature.

In this paper, we show that true 3D placement approaches, enhanced with reinforcement learning, can offer further PPA improvements over pseudo-3D approaches. To accomplish this goal, we integrate an academic true 3D placement engine into a commercial-grade 3D physical design flow, creating ART-3D flow (Analytical 3D Placement with Reinforced Parameter Tuning-based 3D flow). We use a reinforcement learning (RL) framework to find optimized placement parameter settings of the true 3D placement engine for a given netlist and perform high-quality 3D placement. We then use an efficient 3D optimization and routing engine based on a commercial place and route (P&R) tool to maintain or improve the benefits reaped from true 3D placement till design signoff. We evaluate our 3D flow by designing several gate-only and processor benchmarks on a commercial 28nm technology node. Our proposed 3D flow involving true 3D placement offers the best PPA results compared to existing 3D P&R flows and reduces power consumption by up to 31%, improves effective frequency by up to 25%, and therefore reduces power-delay product by up to 43% compared with commercial 2D IC design flow. These improvements predominantly come from RL-based parameter tuning, as it improves the performance of the 3D placer by up to 12%.

A 3D Integrated Circuit consists of two or more dies bonded to each other in the vertical direction. This allows for a high transistor density without a need for shrinking the underlying transistor dimensions. While it has been shown to improve design power, performance, and area (PPA) due to the stacked Front End Of the Line (FEOL) layers, the Back End Of the Line (BEOL) structure of the stacked IC also allows for novel routing scenarios. With the split dies in 3D, nets would need to connect cells from different tiers, across many vertical layers and multiple FEOLs. More importantly, nets connecting cells in a single tier could still use metal layers from the BEOL of other tiers to complete routing. This is referred to as routing / metal layer sharing. While such sharing creates additional 3D connections, it can also be utilized to improve several aspects of the design such as cost, routing congestion, and performance. In this paper, we analyze the nets with metal layer sharing in 3D and provide ways to control the number of 3D connections. We show that the configuration of the 3D BEOL stack helps with metal layer cost reduction with up to 1-2 fewer layers needed to complete routing without a noticeable timing impact. Sharing also allows for a better distribution of wirelength in the BEOL stack that can achieve significant reduction in metal layer congestion of top most layer by up to a 50% reduction of its track usage. Finally, we also see performance benefits of up to 16% with the help of metal layer sharing in 3D IC design.

Monolithic 3D (M3D) integration provides massive vertical integration through the use of nanoscale inter-layer vias (ILVs). However, high integration density and aggressive scaling of the inter-layer dielectric make ILVs especially prone to defects. We present a low-cost built-in self-test (BIST) method that requires only two test patterns to detect opens, stuck-at faults, and bridging faults (shorts) in ILVs. We also propose an extended BIST architecture for fault detection, called Dual-BIST, to guarantee zero ILV fault masking due to single BIST faults and negligible ILV fault masking due to multiple BIST faults. We analyze the impact of coupling between adjacent ILVs arranged in a 1D array in block-level partitioned designs. Based on this analysis, we present a novel test architecture called Shared-BIST with the added functionality of localizing single and multiple faults, including coupling-induced faults. We introduce a systematic clustering-based method for designing and integrating a delay bank with the Shared-BIST architecture for testing small-delay defects in ILVs with minimal yield loss. Simulation results for four two-tier M3D benchmark designs highlight the effectiveness of the proposed BIST framework.

This work addresses how to naturally adopt the l 2 -norm cosine similarity in the neuromemristive system and studies the unsupervised learning performance on handwritten digit image recognition. Proposed architecture is a two-layer fully connected neural network with a hard winner-take-all (WTA) learning module. For input layer, we propose single-spike temporal code that transforms input stimuli into the set of single spikes with different latencies and voltage levels. For a synapse model, we employ a compound memristor where stochastically switching binary-state memristors connected in parallel, which offers a reliable and scalable multi-state solution for synaptic weight storage. Hardware-friendly synaptic adaptation mechanism is proposed to realize spike-timing-dependent plasticity learning. Input spikes are sent out through those memristive synapses to each and every integrate-and-fire neuron in the fully connected output layer, where the hard WTA network motif introduces the competition based on cosine similarity for the given input stimuli. Finally, we present 92.64% accuracy performance on unsupervised digit recognition with only single-epoch MNIST dataset training via high-level simulations, including extensive analysis on the impact of system parameters.

We present a ReRAM memory compiler for monolithic 3D (M3D) integrated circuits (IC). We develop ReRAM architectures for M3D ICs using 1T-1R bit cells and single and multiple tiers of transistors for access and peripheral circuits. The compiler includes an automated flow for generation of subarrays of different dimensions and larger arrays of a target capacity by integrating multiple subarrays. The compiler is demonstrated using an M3D process design kit (PDK) based on a Carbon Nanotube Transistor technology. The PDK includes multiple layers of transistors and back-end-of-the-line integrated ReRAM. Simulations show the compiled ReRAM macros with multiple tiers of transistors reduces footprint and improves performance over the macros with single-tier transistors. The compiler creates layout views that are exported into library exchange format or graphic data system for full-array assembly and schematic/symbol views to extract per-bit read/write energy and read latency. Comparison of the proposed M3D subarray architectures with baseline 2D subarrays, generated with a custom-designed set of bit cells and peripherals, demonstrate up to 48% area reduction and 13% latency improvement.

Monolithic 3D (M3D) is an emerging heterogeneous integration technology that overcomes the limitations of the conventional through-silicon-via (TSV) and provides significant performance uplift and power reduction. However, the ultra-dense 3D interconnects impose significant challenges during physical design on how to best utilize them. Besides, the unique low-temperature fabrication process of M3D requires dedicated design-for-test mechanisms to verify the reliability of the chip. In this article, we provide an in-depth analysis on these design and test challenges in M3D. We also provide a comprehensive survey of the state-of-the-art solutions presented in the literature. This article encompasses all key steps on M3D physical design, including partitioning, placement, clock routing, and thermal analysis and optimization. In addition, we provide an in-depth analysis of various fault mechanisms, including M3D manufacturing defects, delay faults, and MIV (monolithic inter-tier via) faults. Our design-for-test solutions include test pattern generation for pre/post-bond testing, built-in-self-test, and test access architectures targeting M3D.

3-D ICs can continue to improve power, performance, area, and cost beyond traditional Moore’s law scaling limitations by leveraging the third dimension and short vertical interconnects. Several recent studies present methodologies to implement 3-D ICs, but most of these studies implement each tier separately after partitioning a design into multiple tiers, resulting in inaccurate buffer insertion, which becomes more severe in advanced technology nodes. In this article, we present a new methodology called “Cascade2D flow” which utilizes design and microarchitecture insight for tier partitioning and implements 3-D ICs using 2-D commercial tools. By modeling vertical interconnects with sets of anchor cells and dummy wires, Cascade2D flow places, and routes and optimizes multiple tiers simultaneously in the 2-D version of a 3-D IC called “cascade2D design,” which enables accurate buffer insertion. Two flavors of 3-D ICs—monolithic 3-D (M3D) and face-to-face-bonded (F2F-bonded) 3-D ICs—of a commercial in-order, 32-bit application processor at foundry 28 nm, 14/16 nm, and predictive 7-nm technology nodes are implemented using this new methodology. We investigate the power, performance and area improvements of 3-D ICs over the 2-D counterparts to examine the efficacy of the methodology. Our new methodology outperforms the state-of-the-art 3-D IC design flows in the both flavors of 3-D ICs with up to$4\times $better power savings. In the best case, 3-D ICs from Cascade2D flow show 25% better performance at iso-power and 20% lower power at iso-performance.

Modern physical design flows highly depend on design space exploration to find the commercial tools’ clock tree synthesis (CTS) parameters that lead to optimized clock trees. However, such exploration is often time-consuming and computationally inefficient. In this article, we overcome this drawback by proposing a novel framework named GAN-CTS, which utilizes conditional generative adversarial network (GAN) to predict and optimize CTS outcomes. Our framework is built upon three sequential learning stages. First, to precisely characterize distinct designs, we leverage transfer learning to extract netlist features directly from placement images. Second, we perform regression learning using various methods to predict the target CTS outcomes and demonstrate that the proposed multitask learning approach achieves better accuracy than the meta-modeling method adopted by previous works. To fully benefit from the predictions made by our framework, we further quantitatively interpret the importance of each CTS input parameter subject to various design objectives through attribution-based learning. Finally, generative adversarial learning is leveraged to optimize the target clock metrics with the guidance provided by the pretrained regression model. To substantiate the generality of our framework, we perform validations on four unseen netlists that are not utilized in the training process. The experimental results conducted on real-world designs demonstrate that our framework: 1) achieves an average prediction error of 3%; 2) improves the commercial tool’s auto-generated clock tree by 20.7% in clock power, 21.5% in clock wirelength, 36.1% in the worst skew; and 3) reaches an F1-score of 0.93 in the classification task of determining successful and failed CTS runs.

Tier partitioning is one of the most critical stages in monolithic 3-D (M3D) integrated circuits (ICs) implementation flows. It transforms 2-D netlists into 3-D by performing tier assignment for each design instance, which directly impacts the power, performance, and area (PPA) metrics of final 3-D full-chip designs. However, the current state-of-the-art tier partitioning approach named bin-based min-cut algorithm has fundamental flaws that lead to severe drawbacks, such as timing degradation, 3-D routing overhead, and redundant monolithic intertier vias (MIVs) insertion. To overcome these issues, in this article, we propose TP-GNN, an unsupervised graph learning-based tier partitioning framework that utilizes graph neural networks (GNNs) and advanced machine learning (ML) techniques to perform tier partitioning. The proposed framework comprehends design- and technology-related parameters properly so that it is generalizable to various netlists and technologies. In addition, it can be integrated with any style of M3D design flows that require tier assignments of standard cells. In the experiments, we validate the proposed framework on seven industrial designs with two different fashions of M3D implementation flows: 1) partitioning-first (Snap3D) and 2) partitioning-last (Shrunk2D and Compact2D) flows. We demonstrate that our framework, TP-GNN, significantly improves the 3-D quality of results (QoR) across most testing designs in a large margin compared with the bin-based min-cut tier partitioning algorithm. Specifically, in OpenPiton, an RISC-V-based multicore system, we observe 27.4%, 7.7%, and 20.3% improvements in performance, wirelength, and energy-per-cycle, respectively. Finally, we perform a case study by applying the proposed framework to a heterogeneous M3D design flow, Pin3D, on a commercial CPU design and observe that TP-GNN reaches better partitioning solutions than the existing partitioning approaches for heterogeneous 3-D ICs.

As Moore’s law with traditional process node scaling is slowing down, other techniques are required for the advancement of process nodes. In this work, we focus on one such alternative: 3-D physical design of integrated circuits (ICs). While many recent studies have shown the benefits of 3-D IC design on timing and power consumption of circuits, routing in 3-D is solely done with the automatic commercial routers and has not been well studied. In this article, we discuss the various routing scenarios that arise from cell partitioning and the metal layer stack in 3-D. Unlike a 2-D IC, the metal layer configuration in 3-D depends on the orientation in which the dies are bonded together. Due to this, depending on the configuration, cells in one tier tend to use routing layers from the other tier. This is referred to as metal layer (or) routing sharing. This depends on the metal layer stack and the cell partitioning in 3-D, as well as the via pitch used for 3-D connections. By analyzing metal layer sharing in detail, we see that it can help reduce metal layer costs in 3-D as well as improve the power consumption and, in some cases, the maximum achievable performance of the circuits. Overall, the 3-D metal layer cost can decrease by 9% along with an improved power delay product of up to 7.5% just from the routing sharing in monolithic 3-D ICs.

In this paper, we present three commercial-grade 3D IC designs based on state-of-the-art design technologies, specifically micro-bumping (3D die stacking), hybrid bonding (wafer-on-wafer bonding) and monolithic 3D IC (M3D). To highlight trade-offs present in these three designs, we perform analyses on power, performance, and area and the clock tree. We also model the tier-to-tier interconnection in each 3D IC methodology and analyze signal integrity to assess the reliability of each design. From our experiments, hybrid bonding design shows the best timing improvement of 81.4% when compared to its 2D counterpart, while micro-bumping shows the best reliability among 3D IC designs.

Gate sizing for timing optimization is performed extensively throughout electronic design automation (EDA) flows. However, increasing design sizes and time-to-market pressure force EDA tools to maintain pseudo-linear complexity, thereby limiting the global exploration done by the underlying sizing algorithms. Furthermore, high-performance low-power designs are pushing the envelope on power, performance and area (PPA), creating a need for last mile PPA closure using more powerful algorithms. Reinforcement learning (RL) is a disruptive paradigm that achieves high-quality optimization results beyond traditional algorithms. In this paper, we formulate gate sizing as an RL process, and propose RL-Sizer, an autonomous gate sizing agent, which performs timing optimization in an unsupervised manner. In the experiments, we demonstrate that RL-Sizer can improve the native sizing algorithms of an industry-leading EDA tool, Synopsys IC-Compiler II (ICC2), on 6 commercial designs in advanced process nodes (5 – 16nm). RL-Sizer delivers significantly better total negative slack (TNS) and number of violating endpoints (NVEs) on 4 designs with negligible power overhead, while achieving parity on athe others.

In this paper, we present a novel heterogeneous design of Monolithic 3D ICs along with crucial design flow enhancements and better partitioning methods. The heterogeneous M3D ICs are designed with a combination of low-cost, low-power, and low-performance cells on one die and a higher-cost, power, and performance technology variant on the tier, for heterogeneity. These heterogeneous designs out-perform most 2D, 3D variants in Power-Delay Product and Cost metrics. Using 4 different netlists, we see up-to 23% improvement in Performance per Cost, and 16% improvement of Power Delay Product with heterogeneous M3D compared to the best 2D designs.

One of the advantages of 3D IC technology is its ability to integrate different devices such as CMOS, SRAM, and RRAM, or multiple technology nodes of single or different devices onto a single chip due to the presence of multiple tiers. This ability to create heterogeneous 3D ICs finds a wide range of applications, from improving processor performance by integrating better memory technologies to building compute-in-memory ICs to support advanced machine learning algorithms. This paper discusses the current trends and future directions for the physical design of heterogeneous 3D ICs. We summarize various physical design and optimization flows, integration techniques, and existing academic works on heterogeneous 3D ICs.

Memory-bound sparse gathering, caused by irregular random memory accesses, has become an obstacle in several on-demand applications such as embedding lookup in recommendation systems. To reduce the amount of data movement, and thereby better utilize memory bandwidth, previous studies have proposed near-data processing (NDP) solutions. The issue of prior work, however, is that they either minimize data movement effectively at the cost of limited memory parallelism or try to improve memory parallelism (up to a certain degree) but cannot successfully decrease data movement, as prior proposals rely on spatial locality (an optimistic assumption) to utilize NDP. More importantly, neither approach proposes a solution for gathering data from random memory addresses; rather they just offload operations to NDP. We propose an effective solution for sparse gathering, an efficient near-memory intelligent reduction (Fafnir) tree, the leaves of which are all the ranks in a memory system, and the nodes gradually apply reduction operations while data is gathered from any rank. By using such an overall tree, Fafnir does not rely on spatial locality; therefore, it minimizes data movement by performing entire operations at NDP and fully benefits from parallel memory accesses in parallel processing at NDP. Further, Fafnir offers other advantages such as using fewer connections (because of the tree topology), eliminating redundant memory accesses without using costly and less effective caching mechanisms, and being applicable to other domains of sparse problems such as scientific computing and graph analytics. To evaluate Fafnir, we implement it on an XCVU9P Xilinx FPGA and in 7 nm ASAP ASIC. Fafnir looks up the embedding tables up to 21.3× more quickly than the state-of-the-art NDP proposal. Furthermore, the generic architecture of Fafnir allows running classic sparse problems using the same 1.2 mm2hardware up to 4.6× more quickly than the state of the art.

Carbon nanotube FETs (CNFETs) are emerging as an alternative to silicon devices for next-generation computing systems. However, imperfect carbon nanotube deposition during CNFET fabrication can lead to the formation of difficult-to-etch CNT aggregates in the active layer. These CNT aggregates can form parasitic CNFETs (para-FETs) that are modulated by adjoining gate contacts or back-end-of-line metal layers, thereby forming conditional shorts and stuck-at faults. We show that even weak (parametric) para-FETs can lead to a degraded static noise margin in CNFET-based design. We propose ParaMitE, a layout optimization method that horizontally flips selected standard cells in situ to minimize the number of para-FETs that can arise due to unetched CNTs. As we modify only the cell orientation (and not the cell placement), the impact on the power, timing, and wire length of the CNFET-based design is negligible. Simulation results for several benchmarks show that the proposed method can mitigate up to 60% of the possible para-FET locations (90% of the most critical locations) with only a 3% increase in the total wire length. ParaMitE can enable yield ramp-up at the foundry by providing guidance on which para-FETs can be avoided by design, and conversely, which CNT aggregates must be removed through processing steps.

Computer-aided design (CAD) tools traditionally optimize for power, performance, and area (PPA). However, given a vast number of hardware security threats, we call for secure-by-design CAD flows, to adopt principles of secure hardware design and streamline security closure throughout the flow. The stakes are high for integrated circuit (IC) vendors and design companies, as security risks that are not addressed during design will inevitably be exploited in the field, where vulnerabilities are almost impossible to fix. This paper highlights the need for security closure of physical layouts because efforts taken toward securing ICs at higher abstraction layers may be futile without support for securing the tape-out ready layouts.

Modern designs are increasingly reliant on physical design (PD) tools to derive full technology scaling benefits of Moore's Law. Designers often perform power, performance, and area (PPA) exploration through parallel PD runs with different tool configurations. Efficient exploration of PPA is mission-critical for chip designers who are working with stringent time-to-market constraints and finite compute resources. Therefore, a framework that can accurately predict a “doomed run” (i.e., will not meet the PPA targets) at early phases of the PD flow can provide a significant productivity boost by enabling early termination of such runs. Multiple QoR metrics can be leveraged to classify successful or doomed PD runs. In this paper, we specifically focus on the aspect of timing, where our goal is to identify the PD runs that cannot achieve end-of-flow timing results by predicting the post-route total negative slack (TNS) values in early PD phases. To achieve our goal, we develop an end-to-end machine learning (ML) framework that performs TNS prediction by modeling PD implementation as a sequential flow. Particularly, our framework leverages graph neural networks (GNNs) to encode netlist graphs extracted from various PD phases, and utilize long short-term memory (LSTM) networks to perform sequential modeling based on the GNN-encoded features. Experimental results on seven industrial designs with 5:2 train/test split ratio demonstrate that our framework predicts post-route TNS values in high fidelity within 5.2% normalized root mean squared error (NRMSE) in early design stages (e.g., placement, CTS) on the two validation designs that are unseen during training.

In this paper, we present a power, performance, area and cost (PPAC) analysis for large-scale 3D processor designs based on wafer-to-wafer bonding. From the evaluation of our cost model, we investigate a typically disregarded opportunity in 3D that is area savings due to buffer savings and better routability, offering unexpected cost savings. We explore the viability of this factor with the feedback of a state-of-the-art 3D memory-on-logic implementation flow. We show how this affects the PPAC of full-chip GDS implementations of a large-scale manycore processor design. Experiments show that our memory-on-logic 3D implementation offers 7% silicon area savings, resulting in 53.5% footprint reduction. We also obtain a 40% power-performance-cost improvement compared with 2D counterparts

3D integration is becoming a cost-effective way to incorporate more CPU cores and memory to improve the performance of computing systems. Meanwhile, due to the higher power density, power delivery and thermal issues become more significant in multi-tier 3DICs. In this paper, we explore and evaluate multiple design options for an Arm Neoverse-based 3D architecture focusing on power and thermals at 7nm process and sub-10$\mu $m pitch. Using a rapid voltage-drop and thermal analysis methodology, we model a system with a 32-core CPU layer and up to 4 layers of system-level caches, and quantity the trade-offs between performance, cost, voltage-drop, and temperature. A 3-layer configuration shows a good balance with 17% IPC gain and 17% lower cost, while incurring 15mV worse voltage drop and 8.5°C higher temperature compared with 2D. Our studies suggest that the co-optimization of system architecture, technology, and physical design is key for high-performance 3D systems.

Placement is one of the most crucial problems in modern Electronic Design Automation (EDA) flows, where the solution quality is mainly dominated by on-chip interconnects. To achieve target closures, designers often perform multiple placement iterations to optimize key metrics such as wirelength and timing, which is highly time-consuming and computationally inefficient. To overcome this issue, in this paper, we present a graph learning-based framework named PL-GNN that provides placement guidance for commercial placers by generating cell clusters based on logical affinity and manually defined attributes of design instances. With the clustering information as a soft placement constraint, commercial tools will strive to place design instances in a common group together during global and detailed placements. Experimental results on commercial multi-core CPU designs demonstrate that our framework improves the default placement flow of Synopsys IC Compiler II (ICC2) by 3.9% in wirelength, 2.8% in power, and 85.7% in performance.

The state-of-the-art Monolithic 3D (M3D) IC design methodologies~\citem3d:Ku-tcad-Compact2D, m3d:Panth-tcad-Shrunk2D use commercial electronic design automation tools built for 2D ICs to implement a pseudo-3D design and split it into two dies that are routed independently to create an M3D design. Therefore, an accurate estimation of 3D wire parasitics at the pseudo-3D stage is important to achieve a well optimized M3D design. In this paper, we present a regression model based on boosted decision tree learning to better predict the 3D wire parasitics (RCs) at the pseudo-3D stage. Our model is trained using individual net features as well as the full-chip design metrics using multiple instantiations of 8 different netlists and is tested on 3 unseen netlists. Compared to the Compact-2D~\citem3d:Ku-tcad-Compact2D flow on its own as the reference pseudo-3D, the addition of our predictive model achieves up to $2.9 \times$ and $1.7 \times$ smaller root mean square error in the resistance and capacitance predictions respectively. On an unseen netlist design, we observe that our model provides 98.6% and 94.6% RC prediction accuracy in 3D and up to $6.4 \times$ smaller total negative slack of the design compared to the result of Compact-2D flow resulting in a more timing-robust M3D IC. This model is not limited to Compact-2D, and can be extended to other pseudo-3D flows.

3D integration technology is one of the leading options that can advance Moore's Law beyond conventional scaling. Due to the absence of commercial 3D placers and routers, existing 3D physical design flows rely heavily on 2D commercial tools to handle 3D IC physical synthesis. Specifically, these flows build 2D designs first and then convert them into 3D designs. However, several works demonstrate that design qualities degrade during this 2D-3D transformation. In this paper, we overcome this issue with our Snap-3D, a constraint-driven placement approach to build commercial-quality 3D ICs. Our key idea is based on the observation that if the standard cell height is contracted by one half and partitioned into multiple tiers, any commercial 2D placer can place them onto the row structure and naturally achieve high-quality 3D placement. This methodology is shown to optimize power, performance, and area (PPA) metrics across different tiers simultaneously and minimize the aforementioned design quality loss. Experimental results on 7 industrial designs demonstrate that Snap-3D achieves up to 5.4% wirelength, 10.1% power, and 92.3% total negative slack improvements compared with state-of-the-art 3D design flows.

The emerging heterogeneous 3D integration technologies provide a promising solution to improve the performance of electronic systems in the post-Moore era, but the lack of design automation solutions and the challenges in physical design are hindering the applications of these technologies. In this paper, we discuss multiple types and levels of heterogeneous integration enabled by the high-density 3D technologies. We investigate each physical implementation stage from technology setup to placement and routing, identify the design challenges proposed by heterogeneous 3D integration. This paper provides a comprehensive survey on the state-of-the-art physical design methodologies to address these challenges.

Studies have shown that monolithic 3D ( M3D ) ICs outperform the existing through-silicon-via ( TSV ) -based 3D ICs in terms of power, performance, and area ( PPA ) metrics, primarily due to the orders of magnitude denser vertical interconnections offered by the nano-scale monolithic inter-tier vias. In order to facilitate faster industry adoption of the M3D technologies, physical design tools and methodologies are essential. Recent academic efforts in developing an EDA algorithm for 3D ICs, mainly targeting placement using TSVs, are inadequate to provide commercial-quality GDS layouts. Lately, pseudo-3D approaches have been devised, which utilize commercial 2D IC EDA engines with tricks that help them operate as an efficient 3D IC CAD tool. In this article, we provide thorough discussions and fair comparisons (both qualitative and quantitative) of the state-of-the-art pseudo-3D design flows, with analysis of limitations in each design flow and solutions to improve their PPA metrics. Moreover, we suggest a hybrid pseudo-3D design flow that achieves both benefits. Our enhancements and the inter-mixed design flow, provide up to an additional 26% wirelength, 10% power consumption, and 23% of power-delay-product improvements.

Monolithic 3-D (M3-D) integration is an emerging technology that offers significant power, performance, and area benefits for an integrated circuit (IC) design. However, a problem with the 3-D power distribution network in such ICs is that it can lead to high power supply noise (PSN) during the capture cycles in at-speed scan testing for transition delay faults. Therefore, the failure of good chips (i.e., yield loss) resulting from the PSN-induced voltage droop is a major concern for M3-D designs. In this article, we first assess the PSN and voltage droop problems and their impact on path delays for at-speed testing of benchmark M3-D designs. Next, we present an analysis framework to identify test patterns that are most likely to lead to yield loss. We describe a test-pattern reshaping solution based on integer linear programming to make appropriate changes to the test patterns that cause yield loss. Simulation results for four M3-D benchmarks highlight the effectiveness of the proposed solution.

The 2-D CMOS process technology scaling may have reached its pinnacle, yet it is not feasible to manufacture all computing elements at lower technological nodes. This has opened a new branch of chip designing that allows chiplets on different technological nodes to be integrated into a single package using interposers, the passive interconnection mediums. However, establishing a high-frequency communication over an entirely passive layer is one of the significant design challenges of 2.5-D systems. In this article, we present a robust clocking architecture for a 2.5-D system consisting of 64 processor cores. This clocking scheme consists of two major components, namely, interposer clocking and on-chiplet clocking. The interposer clocking consists of clocks used to achieve global synchronicity and clocks for interchiplet communication established using the AIB protocol. We synthesized these clocking components using commercial EDA tools and analyzed them using standard tools, on-chip, and package models. We also compare these results against a 2-D design of the same benchmark and another 2.5-D clocking architecture. Our experiments show that the absolute clock power is up to 16% less, and the ratio of clock power to system power is up to 4% less in the 2.5-D design than its 2-D counterpart.

Resistive random access memory (RRAM)-based compute-in-memory architecture helps overcome the bottleneck caused by large memory transactions in the convolutional neural network (CNN) accelerators. However, their deployment using 2-D IC technology faces challenges, as today's RRAM cells remain at legacy nodes above 20 nm due to high programming voltages. Besides, power-hungry analog-to-digital converter (ADC) units limit the throughput of RRAM accelerators. In this article, we present the first-ever heterogeneous (multiple technology nodes) mixed-signal monolithic 3-D IC designs of the RRAM CNN accelerator. Our RRAM remains at legacy 40-nm nodes in one tier, but CMOS periphery scales toward advanced 28/16 nm in another tier. Our 3-D designs overcome the bottleneck caused by ADCs and offer up to 4.9× improvement in energy efficiency in TOPS/W and up to 50% reduction in footprint area over 40-nm 2-D IC designs. Compared with existing 2-D works, our 3-D architecture offers up to 28.6× improvement in energy efficiency.

Monolithic 3-D IC (M3-D) is a promising solution to improve the performance and energy-efficiency of modern processors. But, designers are faced with challenges in design tools and methodologies, especially for power and thermal verifications. We developed a new physical design flow that optimally places and routes cache modules in one tier and logic gates in the other. Our tool also builds high-quality clock and power delivery networks targeting logic-on-memory M3-D designs. Finally, we developed a sign-off analysis tool flow to evaluate power, performance, area (PPA), thermal, and voltage-drop quality for given M3-D designs. Using our complete register transfer level (RTL)-to-Graphic Design System (GDS) tool flow, we designed commercial quality 2-D and M3-D implementation of Arm Cortex-A7 and Cortex-A53 processors in a commercial 28-nm technology. Experimental results show that our 3-D processors offer 20% (A7) and 21% (A53) performance gain, compared with their 2-D commercial counterparts. The voltage-drop degradation of our 3-D Cortex-A7 and Cortex-A53 processors is less than 3% of the supply voltage, while temperature increase is 10.71 °C and 13.04 °C, respectively.

Monolithic 3D (M3D) integration is an emerging technology that offers significant power, performance, and area benefits for integrated circuit (IC) design. However, a problem with the 3D power distribution network in such ICs is that it can lead to high power supply noise (PSN) during the capture cycles in at-speed scan testing for transition delay faults. Therefore, the failure of good chips (i.e., yield loss) resulting from the PSN-induced voltage droop is a major concern for M3D designs. In this paper, we first assess the PSN and voltage droop problems, and their impact on path delays for at-speed testing of benchmark M3D designs. Next, we present an analysis framework to identify test patterns that are most likely to lead to yield loss. We describe a test-pattern reshaping solution based on integer linear programming to make appropriate changes to the test patterns that cause yield loss. Simulation results for four M3D benchmarks highlight the effectiveness of the proposed solution.

3D integration technology is one of the few options that can keep Moore's Law trajectory beyond conventional scaling. Existing 3D physical design flows fail to benefit from the full advantage that 3D integration provides. Particularly, current 3D partitioning algorithms do not comprehend technology and design-related parameters properly, which results in sub-optimal partitioning solutions. In this paper, we propose TP-GNN, an unsupervised graph-learning-based tier partitioning framework, to overcome this issue. Experimental results on 7 industrial designs demonstrate that our framework significantly improves the QoR of the state-of-the-art 3D implementation flows. Specifically, in OpenPiton, a RISC-V-based multi-core system, we observe 27.4%, 7.7% and 20.3% improvements in performance, wirelength, and energy-per-cycle respectively.

Memory-on-logic and sensor-on-logic face-to-face stacking are emerging design approaches that promise a significant increase in the performance of modern systems-on-chip at reasonable costs. In this work, a netlist-to-layout design flow for such heterogeneous 3D systems is proposed. The proposed technique overcomes the severe limitations of existing 3D physical design methodologies. A RISC-V-based multi-core system, implemented in a commercial technology, is used as a case study to evaluate the proposed design flow. The case study is performed for modern/large and small cache sizes to show the superiority of the proposed methodology for a broad set of systems. While previous 3D design flows do not show to optimize performance against 2D baseline designs for processor systems with a significant memory area occupation, the proposed flow shows a performance and power improvement by 20.4-28.2% and 3.2-3.8%, respectively.

The quality of placement is essential in the physical design flow. To achieve PPA goals, a human engineer typically spends a considerable amount of time tuning the multiple settings of a commercial placer (e.g. maximum density, congestion effort, etc.). This paper proposes a deep reinforcement learning (RL) framework to optimize the placement parameters of a commercial EDA tool. We build an autonomous agent that learns to tune parameters optimally without human intervention and domain knowledge, trained solely by RL from self-search. To generalize to unseen netlists, we use a mixture of handcrafted features from graph topology theory along with graph embeddings generated using unsupervised Graph Neural Networks. Our RL algorithms are chosen to overcome the sparsity of data and latency of placement runs. Our trained RL agent achieves up to 11% and 2.5% wirelength improvements on unseen netlists compared with a human engineer and a state-of-the-art tool auto-tuner, in just one placement iteration (20× and 50× less iterations).

In this paper, we propose RTL-to-GDS design flow for monolithic 3D ICs (M3D) built with carbon nanotube field-effect transistors and resistive memory. Our tool flow is based on commercial 2D tools and smart ways to extend them to conduct M3D design and simulation. We provide a post-route optimization flow, which exploits the full potential of the underlying M3D process design kit (PDK) for power, performance and area (PPA) optimization. We also conduct IR-drop and thermal analysis on M3D designs to improve the reliability. To enhance the testability of our M3D designs, we develop design-for-test (DFT) methodologies and integrate a low-overhead built-in self-test module into our design for testing inter-layer vias (ILVs) as well as logic circuitries in the individual tiers. Our benchmark design is RISC-V Rocketcore, which is an open source processor. Our experiments show 8.1% of power, 19.6% of wirelength and 55.7% of area savings with M3D designs at iso-performance compared to its 2D counterpart. In addition, our IR-drop and thermal analyses indicate acceptable power and thermal integrity in our M3D design.

Modern high-performance System-on-Chip (SoC) design flows highly depend on signoff tools to perform timing-constrained power optimization through Engineering Change Orders (ECOs), which involve gate-sizing and Vth-assignment of standard cells. However, ECOs are highly time-consuming, and the power improvement is unknown in advance. Ever since the industrial benchmarks released by the ISPD-2012 gate-sizing contest, active research has been conducted extensively. Nonetheless, previous works were mostly based on heuristics or analytical methods whose timing models were over-simplified and lacked formal validations from commercial signoff tools. In this paper, we propose ECO-GNN, a transferable graph-learning-based framework, which harnesses graph neural networks (GNNs) to perform commercial-quality signoff power optimization through discrete Vth-assignment. Our framework generates tool-accurate optimization results instantly on unseen netlists that are not utilized in the training process. Furthermore, we implement a GNN-based explanation method to interpret the optimization results achieved by our framework. Experimental results on 14 industrial designs, including a RISC-V-based multi-core system and the renowned ISPD-2012 benchmarks, demonstrate that our framework achieves up to 14X runtime improvement with similar signoff power optimization quality compared with Synopsys PrimeTime.

In this paper, we present an optimization flow for monolithic 3D ICs called Pin-3D Optimizer. Compared with the state-of-the-art RTL-to-GDS flows that rely on ad-hoc technology file tweaks and RC scaling, Pin-3D offers a streamlined method to run commercial 2D IC tools to obtain high-quality monolithic 3D IC designs. Specifically, Pin-3D supports effective legalization, routing, timing closure, and ECO optimization for monolithic 3D IC designs. We propose a novel optimization methodology where the cells in each tier of a 3D IC are optimized using cell data and constraints of the full 3D design. The optimizations in a tier also directly influence the timing, power in the other tiers, leading to better overall PPA of the 3D IC. With the help of two industry processors designed with a 28 nm technology node, we show that Pin-3D provides up-to 9.0% smaller wirelength and 88% smaller total negative slack than die-by-die M3D flows. We also observe up-to 8.7% lower power and 26% smaller wirelength than 2D ICs. In addition, Pin-3D is the first flow that supports routing and timing optimization in heterogeneous logic-on-logic monolithic 3D ICs. We demonstrate this capability by performing area-balanced tier partitioning, routing, and timing closure of a 3D design with different technologies on each die.

In this work, we develop a new 3D flash memory cube architecture that integrates multiple flash dies and their logic controller in a unique and optimized fashion for space applications. In our Loaf-of-Bread (LOB) configuration, flash dies are standing up and bonded laterally instead of the conventional pancake-style vertical die stacking. Our LOB allows the flash dies to be bonded without the use of through-silicon-vias and micro-bumps. Instead, we insert a redistribution layer in between two adjacent dies to bring the IO signals to the bottom, where a logic controller die collects the IOs from all flash dies and coordinates the communication with an off-cube host processor. Thus, our LOB configuration allows the users to form flash memory cubes using off-the-shelf 2D flash dies and complete the integration at a packaging house instead of a fab. A key element in our LOB flash cube is the logic controller architecture that supports fault-tolerant and energy-efficient operation of the cube. We develop the controller architecture and validate the functionality using C++ emulation and FPGA prototyping. Compared with a state-of-the-art space-grade flash memory module, our system shows a 20X bandwidth improvement in a smaller form factor along with a 25X better ratio of density per volume.

The optimal selection of an interposer substrate is important in 2.5D systems, because its physical, material and electrical characteristics govern the overall system performance, reliability and cost. Several materials have been proposed that offer various tradeoffs including silicon, organic, glass and etc. In this paper, we conduct a quantitative comparison between two 2.5D IC designs based on silicon vs. liquid crystal polymer (LCP) interposer technologies in the overall system level for the first time. We also investigate tradeoffs in power, performance and area (PPA), signal integrity (SI) and power integrity (PI) depending on the interposer technologies. Through our flow, we generate a large-scale benchmark architecture with commercial-grade GDS layouts of interposer and chiplets using two different interposer substrates. Then, we model transmission lines and power delivery network (PDN) of each 2.5D IC design. Finally, we perform PPA analysis, SI and PI on both 2.5D IC designs to observe the quantitative tradeoffs between two designs. Our experiment shows that silicon interposer-based design has 10.46% less power, 0.25× smaller area and 0.57× shorter average wirelength compared to LCP interposer-based design. However, LCP-based design has 0.59× smaller PDN DC impedance and 0.75× shorter worst delay of interposer wire while maintaining the power delivery efficiency. Lastly, our cost analysis of 2.5D IC design indicates that the overall cost of organic LCP technology, if both the chiplets and their interposer costs are combined, is 2.69× higher than the silicon even the cost of LCP interposer is 1.91% of silicon interposer. This indicates that LCP technology is prohibitive unless the interconnect and bump dimensions are dramatically reduced.

In a 2D design, the periphery of a block serves as the optimal pin location since blocks are placed aside horizontally in a single placement layer. However, Monolithic 3D (M3D) integration relieves this boundary constraint by allowing vertical block communication between different tiers based on a nm-scale 3D interconnection pitch. In this paper, we present a design methodology named Pin-in-the-Middle that assigns block pins in the middle of a block using commercial 2D P&R tools to enable efficient block implementation and integration for two-tier block-level M3D ICs. Based on a 28nm two-tier M3D hierarchical design result, we show that our solution offers 13.6% and 24.7% energy-delay-product reduction compared to the M3D design with pins assigned at the block boundaries and its 2D counterpart, respectively.

Despite the recent academic efforts to develop Electronic Design Automation (EDA) algorithms for 3D ICs, the current market does not have commercial 3D computer-aided design (CAD) tools. Insteadpseudo-3D alternative design flows have been devised which utilize commercial 2D CAD engines with tricks that help them operate as a fairly-efficient 3D CAD tool. In this paper we provide detailed discussions and fair power-performance-area (PPA) comparisons of state-of-the-art pseudo-3D design flows. We also analyze the limitations of each design flow and provide solutions with better PPA and various design options. Our experiments using commercial PDK, GDS layouts, and sign-off simulations demonstrate that we achieve up to 26% wirelength and 10% power consumption reduction for pseudo-3D design flows. We also provide a partitioning-first scheme to partitioning-last design flow which increases design freedom with tolerable PPA degradation.

Due to the short die-to-die distance and inferior heat dissipation capability, Face-to-Face (F2F) boned 3D ICs are often considered to be vulnerable to electrical and thermal coupling. This study is the first to quantify the impacts of the electro-thermal coupling on the full-chip timing, power, and performance. We first present an implementation flow for realistic F2F 3D ICs including pad layers and power grids. Then, we propose our signal integrity analysis, parasitic extraction, and thermal analysis flows. Next, we investigate the impacts of the coupling on the delay, power, and noise of F2F 3D ICs, and provide guidelines to mitigate these effects. Our experimental results show that the inter-die electrical coupling causes up to 5.81% timing degradation and 4.00% noise increase, while the thermal coupling leads to less than 0.41% timing degradation and nearly no noise increase. The impact of the combined electro-thermal coupling on delay and noise reaches 6.07% and 4.05%, respectively.

Compute in-memory (CIM) is a promising technique that minimizes data transport, the primary performance bottleneck and energy cost of most data intensive applications. This has found wide-spread adoption in accelerating neural networks for machine learning applications. Utilizing a crossbar architecture with emerging non-volatile memories (eNVM) such as dense resistive random access memory (RRAM) or phase change random access memory (PCRAM), various forms of neural networks can be implemented to greatly reduce power and increase on chip memory capacity. However, compute in-memory faces its own limitations at both the circuit and the device levels. Although compute in-memory using the crossbar architecture can greatly reduce data transport, the rigid nature of these large fixed weight matrices forfeits the flexibility of traditional CMOS and SRAM based designs. In this work, we explore the different synchronization barriers that occur from the CIM constraints. Furthermore, we propose a new allocation algorithm and data flow based on input data distributions to maximize utilization and performance for compute-in memory based designs. We demonstrate a$\boldsymbol{7.47}\times$performance improvement over a naive allocation method for CIM accelerators on ResNet18.

The recent advancement of wafer bonding and monolithic integration technology offers fine-grained 3-D interconnections to face-to-face (F2F) and monolithic 3-D (M3D) ICs. In this article, we propose a full-chip RTL-to-GDSII physical design solution to build commercial-quality two-tier gate-level F2F and M3D ICs. The state-of-the-art flow named shrunk2D (S2D) requires shrinking of standard cells and interconnects by a factor of 50% to fit into the target 3-D footprint of a two-tier design. This, unfortunately, necessitates commercial place/route engines that handle one node smaller geometries, which can be challenging and costly. Our flow named compact-2D (C2D) does not require any geometry shrinking. Instead, C2D implements a 2-D IC with scaled interconnect RC parasitics and contracts the layout to the 3-D integrated circuit footprint. In addition, C2D offers post-tier-partitioning optimization (post-TP opt) which is completely missing in S2D. This additional optimization step is shown to be effective in fixing timing violations caused by inter-tier 3-D routing overhead. Lastly, we present a methodology to reuse the routing result of post-TP opt for the final GDSII generation. Our experimental results show that at iso-performance, C2D offers up to 28.0% power reduction and 15.6% silicon area savings over commercial 2-D ICs without any routing resource overhead.

We present a machine learning (ML) framework to improve the use of computing resources in the FPGA compilation step of a commercial FPGA-based logic emulation flow. Our ML models enable highly accurate predictability of the final place and route design qualities, runtime, and optimal mapping parameters. We identify key compilation features that may require aggressive compilation efforts using our ML models. Experiments based on our large-scale database from an industry’s emulation system show that our ML models help reduce the total number of jobs required for a given netlist by 33%. Moreover, our job scheduling algorithm based on our ML model reduces the overall time to completion of concurrent compilation runs by 24%. In addition, we propose a new method to compute “recommendations” from our ML model to perform re-partitioning of difficult partitions. Tested on a large-scale industry system on chip design, our recommendation flow provides additional 15% compile time savings for the entire system on chip. To exploit our ML model inside the time-critical multi-FPGA partitioning step, we implement it in an optimized multi-threaded representation.

The first mainstream products in three-dimensional integrated circuit (3-D IC) design are memory devices where multiple memory tiers are horizontally integrated to offer manifold improvements when compared with their 2-D counterparts. Unfortunately, none of these existing 3-D memory cubes are ready for harsh space environments. This article introduces a new memory cube architecture for space, based on the vertical integration of multiple commercial-off-the-shelf, 3-D stacked, dynamic random-access memory (DRAM) memory devices with a custom radiation-hardened-by-design controller. Our solution offers high memory capacity, increased bandwidth, fault tolerance, and improved size-weight-and-power characteristics needed for space missions. Validation and functional evaluation of the application-specific integrated circuit (ASIC) controller will be conducted prior to tape-out on a custom FPGA-based emulator platform integrating the 3-D stack. The selected test methodology ensures high-quality register transfer level (RTL) as well as allows to subject the cube structure to radiation testing. The proposed design concept allows for flexibility in the choice of the DRAM die in the case of technology road-map changes or unsatisfactory radiation results.

A new trend in system-on-chip (SoC) design is chiplet-based IP reuse using 2.5-D integration. Complete electronic systems can be created through the integration of chiplets on an interposer, rather than through a monolithic flow. This approach expands access to a large catalog of off-the-shelf intellectual properties (IPs), allows reuse of them, and enables heterogeneous integration of blocks in different technologies. In this article, we present a highly integrated design flow that encompasses architecture, circuit, and package to build and simulate heterogeneous 2.5-D designs. Our target design is 64core architecture based on Reduced Instruction Set Computer (RISC)-V processor. We first chipletize each IP by adding logical protocol translators and physical interface modules. We convert a given register transfer level (RTL) for 64-core processor into chiplets, which are enhanced with our centralized network-onchip. Next, we use our tool to obtain physical layouts, which is subsequently used to synthesize chip-to-chip I/O drivers and these chiplets are placed/routed on a silicon interposer. Our package models are used to calculate power, performance, and area (PPA) and reliability of 2.5-D design. Our design space exploration (DSE) study shows that 2.5-D integration incurs 1.29× power and 2.19× area overheads compared with 2-D counterpart. Moreover, we perform DSE studies for power delivery scheme and interposer technology to investigate the tradeoffs in 2.5-D integrated chip (IC) designs.

A new trend in complex SoC design is chiplet-based IP reuse using 2.5D integration. In this paper we present a highly-integrated design flow that encompasses architecture, circuit, and package to build and simulate heterogeneous 2.5D designs. We chipletize each IP by adding logical protocol translators and physical interface modules. These chiplets are placed/routed on a silicon interposer next. Our package models are then used to calculate PPA and signal/power integrity of the overall system. Our design space exploration study using our tool flow shows that 2.5D integration incurs 2.1x PPA overhead compared with 2D SoC counterpart.

Monolithic 3D IC overcomes the limitation of the existing through-silicon-via (TSV) based 3D IC by providing denser vertical connections with nano-scale inter-layer vias (ILVs). In this paper, we demonstrate a thorough RTL-to-GDS design flow for monolithic 3D IC, which is based on commercial 2D place-and-route (P&R) tools and clever ways to extend them to handle 3D IC designs and simulations. We also provide a low-cost built-in-self-test (BIST) method to detect various faults that can occur on ILVs. Lastly, we present a resistive random access memory (ReRAM) compiler that generates memory modules that are to be integrated in monolithic 3D ICs.

Monolithic 3D integration provides massive vertical integration through the use of nanoscale inter-layer vias (ILVs). However, high integration density and aggressive scaling of the inter-layer dielectric make ILVs especially prone to defects. We present a low-cost built-in self-test (BIST) method to detect opens, stuck-at faults (SAFs), and bridging faults (shorts) in ILVs. Two test patterns-all-1s and all-0s-are applied to the input side of a set of ILVs (e.g., making up a bus between two tiers). On the adjacent tier (the output side of the ILVs), the test responses are compacted to a 2-bit signature through space compaction. We prove that this compaction solution does not introduce any fault aliasing. Simulations results using HSPICE and M3D benchmark designs show that the proposed BIST method requires low area overhead and test time, but provides effective fault localization and the detectability of a wide range of resistive faults.

Monolithic 3D (M3D) ICs provide a way to achieve high performance and low power designs within the same technology node, thereby bypassing the need for transistor scaling. M3D ICs have multiple 2D tiers sequentially fabricated on top of each other and connected through Monolithic Inter-tier vias (MIVs) which go from the top metal of the bottom die to the top metal of the top die. MIVs go through the dielectric separation between tiers and the active layer of the top die. The traditional APR tools only support place and route optimization in 2D IC designs and thus cannot handle the M3D designs. This paper provides a survey of all the efforts done to implement commercial quality M3D ICs using 2D APR tools and the benefit of M3D ICs over traditional 2D ICs.

This paper presents a machine learning (ML) framework to improve the use of computing resources in the FPGA compilation step of a commercial FPGA-based logic emulation flow. Our ML models enable highly accurate predictability of the final P&R design qualities, runtime, and optimal mapping parameters. We identify key compilation features that may require aggressive compilation efforts using our ML models. Experiments based on our large-scale database from an industry's emulation system show that our ML models help reduce the total number of jobs required for a given netlist by 33%. Moreover, our job scheduling algorithm based on our ML model reduces the overall time to completion of concurrent compilation runs by 24%. In addition, we propose a new method to compute “recommendations” from our ML model, in order to perform repartitioning of difficult partitions. Tested on a large-scale industry SoC design, our recommendation flow provides additional 15% compile time savings for the entire SoC.

In this paper, we propose a complete framework named GAN-CTS which utilizes conditional generative adversarial network (GAN) and reinforcement learning to predict and optimize clock tree synthesis (CTS) outcomes. To precisely characterize different netlists, we leverage transfer learning to extract design features directly from placement images. Based on the proposed framework, we further quantitatively interpret the importance of each CTS input parameter subject to various design objectives. Finally, to prove the generality of our framework, we conduct experiments on the unseen netlists which are not utilized in the training process. Experimental results performed on industrial designs demonstrate that our framework (1) achieves an average prediction error of 3%, (2) improves the commercial tool's auto-generated clock tree by 51.5% in clock power, 18.5% in clock wirelength, 5.3% in the maximum skew, and (3) reaches an F1-score of 0.952 in the classification task of determining successful and failed CTS processes.

Integrated voltage regulators (IVR) with embedded inductors is an emerging technology that provides point-of-load voltage regulation to high-performance systems. Conventional two-step approaches to the design of IVRs can suffer from suboptimal design as the optimal inductor depends on the characteristics of the buck converter (BC). Furthermore, inductor-level trade-offs such as AC and DC resistance, inductance and area can not be determined independently from the BC. This co-dependency of the BC and the inductor creates a highly non-linear response surface, which raises the necessity of co-optimization, involving multiple time-consuming electromagnetics (EM) simulations. In this paper, we propose a machine learning based optimization methodology that eliminates EM simulations from the optimization loop to significantly reduce the optimization complexity. A novel technique named as Spectral Transposed Convolutional Neural Network (S-TCNN) is presented to derive an accurate predictive model of the inductor frequency response using a small amount of training data. The derived S-TCNN is then used along with a time-domain model of the BC to perform multi-objective optimization that approximates the Pareto front for 5 objectives, namely inductor area, BC settling time, voltage conversion efficiency, droop and ripple. The resulting methodology provides multiple Pareto optimal inductors in an efficient and fully automated fashion, thereby allows to rapidly determine the optimal trade-offs for possibly contradicting design objectives. We demonstrate the proposed framework on co-optimization of solenoidal inductor with magnetic core and BC that are integrated on silicon interposer.

In this paper, we propose simple but effective clock tree optimization algorithms for monolithic 3D ICs that are based on tier partitioning and flip-flop relocation. Our algorithms take into account 3D timing critical paths, clock skew, and the clock tree hierarchy for a better quality 3D clock tree. We also perform clock slew manipulation and buffer reduction to further improve the 3D designs. We tested four industrial benchmarks implemented using a commercial library and observed up to 34.3% clock skew, 35.9% clock wirelength, 10.0% combinational clock power, and 15.5% total power savings compared to the state-of-the-art [8].

Through-silicon vias (TSVs) are subject to thermal fatigue due to stress over time, no matter how small the stress is. Existing works on TSV fatigue all rely on measurement-based parameters to estimate the lifetime, and cannot consider detailed thermal profiles. In this paper, we propose a new method for TSV fatigue prediction using entropy production during thermal cycles, which is validated by theoretical analysis and measurement results. By combining thermodynamics and mechanics laws, the fatigue process can be quantitatively evaluated with detailed thermal profiles. Experimental results show that interestingly, the landing pad possesses the most easy-to-fail region, which generates up to 50% more entropy compared with the TSV body. The impact of landing pad dimension and TSV geometries are also studied, providing guidance for reliability enhancement. Full-chip fatigue analysis is performed based on stress superposition. Based on the developed theory, we study the interesting problem of fatigue lifetime of 3-D microprocessors in smartphones and its dependence on user task profile. Finally the lifetime of 3D IC TSV in smartphone is researched with selected apps and user profiles. To the best of the authors' knowledge, this is the first TSV fatigue model that is free of measurement data fitting, the first that is capable of considering detailed thermal profiles, and the first framework for efficient full-chip TSV fatigue analysis, and the first study on fatigue lifetime of mobile chips.

As 2-D scaling reaches its limit, monolithic 3-D (M3D) IC is a leading contender for continuing equivalent scaling. Although M3D shows power and performance benefits over 2-D designs, designing a power delivery network (PDN) for M3D is challenging. In this paper, for the first time, we present a system-level PDN model of M3D designs focusing on both resistive (IR) and inductive (Ldi/dt) components of power supply integrity. In addition, we present frequency- and time-domain analyses of M3D PDNs. We show that the additional resistance in M3D PDNs, while being worse for resistive drops, improves resiliency against ac current noise showing 35.9% peak impedance reduction compared to 2-D PDNs during worst case resonant oscillations. Then, we present methodologies to improve power supply integrity of M3D designs based on the observations. Our optimization methodologies offer up to 32.6% and 17.0% static and dynamic voltage drop reduction compared to the baseline M3D designs, respectively, showing 9.0% lower dynamic voltage drop compared to the 2-D counterparts.

The first mainstream products in 3D IC design are memory devices where multiple memory tiers are horizontally integrated to offer manifold improvements compared with their 2D counterparts. Unfortunately, none of these existing 3D memory cubes are ready for harsh space environments. This paper presents a new memory cube architecture for space, based on vertical integration of Commercial-Off-The-Shelf (COTS), 3D stacked, DRAM memory devices with a custom Radiation-Hardened-By-Design (RHBD) controller offering high memory capacity, robust reliability and low latency. Validation and evaluation of the ASIC controller will be conducted prior to tape-out on a custom FPGA-based emulator platform integrating the 3D-stack.

A liquid state machine (LSM) is a powerful recurrent spiking neural network shown to be effective in various learning tasks including speech recognition. In this work, we investigate design and architectural co-optimization to further improve the area-energy efficiency of LSM-based speech recognition processors with monolithic 3D IC (M3D) technology. We conduct fine-grained tier partitioning, where individual neurons are folded, and explore the impact of shared memory architecture and synaptic model complexity on the power-performance-area-accuracy (PPAA) benefit of M3D LSM-based speech recognition. In training and classification tasks using spoken English letters, we obtain up to 70.0% PPAA savings over 2D ICs.

Monolithic three-dimensional integration (M3I) directly fabricates tiers of integrated circuits upon each other and provides millions of vertical interconnections with inter-layer vias (ILVs). It thus brings higher integration density and communication capability compared with three-dimensional stacked integration (3D-SI). However, the Known-Good-Die problem haunting 3D-SI-a faulty tier causes the failure of the entire stack-also occurs in M3I. Lack of efficient test methodologies such as the pre-bond testing in 3D-SI, M3I may have a more significant yield drop and thus its cost may be unacceptable for main-stream adoption. This paper introduces a novel In-growth test method for M3I SRAM. We propose a novel Design-for-Test (DfT) methodology to enable the proposed In-growth test on cell-level partitioned incomplete SRAM cells. We also build a statistical model of cost and discover a prospective judgement to determine whether or not to stop the fabrication, in order to prevent from raising the cost of fabricating more tiers upon the irreparable tiers. We find that a “sweet point” exists in the judgement, which can minimize the overall cost. Experimental results show the effectiveness of our proposed test methodology.

As small-form-factor and low-power end devices matter in the cloud networking and Internet-of-Things Era, the bio-inspired neuromorphic architectures attract great attention recently in the hope of reaching the energy-efficiency of brain functions. Out of promising solutions, a liquid state machine (LSM), that consists of randomly and recurrently connected reservoir neurons and trainable readout neurons, has shown a great promise in delivering brain-inspired computing power. In this work, we adopt the state-of-the-art face-to-face (F2F)-bonded 3D IC flow named Compact-2D [4] to the LSM processor design, and study the power-area-accuracy benefits of 3D LSM ICs targeting the next generation commercial-grade neuromorphic computing platforms. First, we analyze how the different size and connection density of a reservoir in the LSM architecture affects the learning performance using the real-world speech recognition benchmark. Also, we explore how much the power-area design overhead should be paid off to enable better classification accuracy. Based on the power-area-accuracy trade-off, we implement a F2F-bonded 3D LSM IC using the optimal LSM architecture, and finally justify that 3D integration practically benefits the LSM processor design in huge form factor and power savings while preserving the best learning performance.

As we approach the limits of 2D device scaling, monolithic 3D IC (M3D) has emerged as a potential solution offering performance and power benefits. Although various studies have been done to increase power savings of M3D designs, efforts to improve their performance are rarely made. In this paper, we, for the first time, perform in-depth analysis of the factors that affect the performance of M3D, and present methodologies to improve the performance. Our methodologies outperform the state-of-the-art M3D design flow by offering 15.6% performance improvement and 16.2% energy-delay product (EDP) benefit over 2D designs.

The recent advancement of wafer bonding technology offers fine-grained and silicon-space overhead-free 3D interconnections in face-to-face (F2F) bonded 3D ICs. In this paper, we propose a full-chip RTL-to-GDSII physical design solution to build high-density and commercial-quality two-tier F2F-bonded 3D ICs. The state-of-the-art flow named Shrunk-2D (S2D) requires shrinking of standard cells and interconnects by a factor of 50% to fit into the target 3D footprint of a two-tier design. This, unfortunately, necessitates commercial place/route engines that handle one node smaller geometries, which can be challenging and costly. Our flow named Compact-2D (C2D) does not require any geometry shrinking. Instead, C2D implements a 2D IC with scaled interconnect RC parasitics, and contracts the layout to the F2F design footprint. In addition, C2D offers post-tier-partitioning optimization that is shown to be effective in fixing timing violations caused by inter-tier 3D routing, which is completely missing in S2D. Lastly, we present a methodology to recycle the routing result of post-tier-partitioning optimization for final GDSII generation. Our experimental results show that at iso-performance, C2D offers up to 26.8% power reduction and 15.6% silicon area savings over commercial 2D ICs without any routing resource overhead.

In recent years, deep learning has become widespread for various real-world recognition tasks. In addition to recognition accuracy, energy efficiency and speed (i.e., performance) are other grand challenges to enable local intelligence in edge devices. In this article, we investigate the adoption of monolithic three-dimensional (3D) IC (M3D) technology for deep learning hardware design, using speech recognition as a test vehicle. M3D has recently proven to be one of the leading contenders to address the power, performance, and area (PPA) scaling challenges in advanced technology nodes. Our study encompasses the influence of key parameters in DNN hardware implementations towards their performance and energy efficiency, including DNN architectural choices, underlying workloads, and tier partitioning choices in M3D designs. Our post-layout M3D designs, together with hardware-efficient sparse algorithms, produce power savings and performance improvement beyond what can be achieved using conventional 2D ICs. Experimental results show that M3D offers 22.3% iso-performance power saving and 6.2% performance improvement, convincingly demonstrating its entitlement as a solution for DNN ASICs. We further present architectural and physical design guidelines for M3D DNNs to maximize the benefits.

Monolithic 3D ICs (M3D) are fabricated using a sequential process that grows new device and interconnect tiers in a bottom-up fashion. This fabrication process is in contrast to through-silicon via (TSV) technology that aligns and bonds pre-built tiers. M3D offers key advantages over TSVs, including (1) orders-of-magnitude smaller inter-tier vias, (2) no need for high alignment accuracy, (3) finer-grained tier partitioning options, etc. Recent studies have shown power, performance, area, and reliability (PPAR) advantages of M3D over TSV. However, M3D also suffers from its own problems, including (1) device and interconnect performance mismatch between tiers, (2) lack of EDA solutions, (3) testing challenges, (4) cost, etc. Research efforts have also been made to model and mitigate the impact of these undesirable characteristics of M3D. We will provide a survey of work that address the above issues and conclude with future directions.

In this paper, we present full-chip designs and their power, performance, and area (PPA) metrics using the ASAP7 process design kit (PDK) and library. Reliable cell library is a key element in evaluating new technological options such as monolithic 3D (M3D) ICs. Given an RTL, we conduct synthesis and place/route to obtain commercial-quality 2D and M3D IC designs and compare PPA. The ASAP7 library is highly useful to build high-quality designs that accurately reflect 7nm technology node. In addition, the full front-end and back-end access provided in ASAP7 allows us to see the impact of various device and interconnect parameters at the full-chip level for both 2D and monolithic 3D ICs. This work demonstrates the critical role of an academic PDK and library in enabling high-quality research in disruptive technologies such as M3D integration.

As 2D scaling reaches its limit, monolithic 3D IC (M3D) is a leading contender to continue equivalent scaling. Although M3D shows power and performance benefits over 2D designs, designing a power delivery network (PDN) for M3D is challenging. In this paper, for the first time, we present a system-level PDN model of M3D designs focusing on both resistive (IR) and inductive (Ldi/dt) components of power-supply integrity. In addition, we present frequency- and time-domain analysis of the M3D PDN. We show that the additional resistance in the M3D PDN, while being worse for resistive drops, improves resiliency against current noise showing 35.9% peak impedance reduction during worst-case resonant oscillations.

In recent years, deep learning has become widespread for various real-world recognition tasks. In addition to recognition accuracy, energy efficiency is another grand challenge to enable local intelligence in edge devices. In this paper, we investigate the adoption of monolithic 3D IC (M3D) technology for deep learning hardware design, using speech recognition as a test vehicle. M3D has recently proven to be one of the leading contenders to address the power, performance and area (PPA) scaling challenges in advanced technology nodes. Our study encompasses the influence of key parameters in DNN hardware implementations towards energy efficiency, including DNN architectural choices, underlying workloads, and tier partitioning choices in M3D. Our post-layout M3D designs, together with hardware-efficient sparse algorithms, produce power savings beyond what can be achieved using conventional 2D ICs. Experimental results show that M3D offers 22.3% iso-performance power saving, convincingly demonstrating its entitlement as a solution for DNN ASICs. We further present architectural guidelines for M3D DNNs to maximize the power saving.

Existing transistor-level monolithic 3D (T-M3D) standard cell layouts are based on the folding scheme, in which the pull-down network is simply folded and placed on top of the pull-up network. In this paper, we propose a new layout method, the stitching scheme, targeted towards improved cell performance and power integrity. We perform extensive analysis on each layout scheme and evaluate the timing/power benefits of the stitching scheme. Since the ground and power rails overlap in the T-M3D layouts with the folding scheme, we also present a design methodology for the power delivery network of folding T-M3D ICs to evaluate the impact of the T-M3D cell layout scheme on static power integrity. Compared to 2D ICs at iso-performance, stitching T-M3D ICs show a maximum of 6% power savings, 44% area savings with only 1% more static IR-drop in the 14nm technology node while folding T-M3D ICs undergo serious degradation in static power integrity, causing a reliability issue.

We study, for the first time, full chip power benefits of negative capacitance FET (NCFET) device technology for commercial-grade GDSII-level designs. Owing to sub-60mV/decade characteristics, NCFETs provide significantly higher drive-current than standard FETs at a given voltage, enabling significant iso-performance power savings by lowering VDD. We use SPICE models of NCFETs corresponding to 14nm node, which incorporate experimentally calibrated models of ferroelectric. We then characterize NCFET-based standard-cell libraries followed by full-chip NCFET-based GDSII-level design implementations of different benchmarks. Our results show that even with increased device capacitance, we can achieve about 4× (up to 74.7%) full-chip power reduction with low-VDD NCFETs over nominal VDD baseline FETs at iso-performance. The power savings are consistent across multiple benchmarks and are higher for low power designs.

Near-threshold computing (NTC) circuits have been shown to offer significant energy efficiency and power benefits but with a huge performance penalty. This performance loss exacerbates if process and voltage variations are considered. In this article, we demonstrate that three-dimensional (3D) IC technology can overcome this limitation. We present a detailed case study with a 28nm commercial-grade core at 0.6V operation optimized with various 3D IC physical design methods. First, our study under the deterministic case shows that 3D IC NTC design outperforms 2D IC NTC by 29.5% in terms of performance at comparable energy. This is significantly higher than the 12.8% performance benefit of 3D IC at nominal voltage supplies due to higher delay sensitivity to input slew at lower voltages. Second, it is well demonstrated that transistor delay is more sensitive to voltage changes at NTC operation. However, our full-chip study reveals that IR drop effect on 2D/3D IC NTC performance is not severe due to the low power consumption and hence lower IR drop values. Third, die-to-die variation impact on full-chip performance is visible in 3D IC NTC designs, but it is not worse compared to 2D IC NTC designs. This is mainly due to the shorter critical path length in 3D IC NTC designs.

Vertically integrated circuits (3-D ICs) may revitalize Moore's law scaling which has slowed down in recent years. 3-D stacking is an emerging technology that stacks multiple dies vertically to achieve higher transistor density independent of device scaling. They provide high-density vertical interconnects, which can reduce interconnect power and delay. Moreover, 3-D ICs can integrate disparate circuit technologies into a single chip, thereby unlocking new system-on-chip architectures that do not exist in 2-D technology. While 3-D integration could bring new architectural opportunities and significant performance enhancement, new thermal, power delivery, signal integrity and reliability challenges emerge as power consumption grows, and device density increases. Moreover, the significant expansion of CPU design space in 3-D requires new architectural models and methodologies for design space exploration (DSE). New design tools and methods are required to address these 3-D-specific challenges. This keynote paper focuses on the state of the art, ongoing advances and future challenges of 3-D IC design tools and methods. The primary focus of this paper is TSV-based 3-D ICs, although we also discuss recent advances in monolithic 3-D ICs. The objective of this paper is to provide a unified perspective on the fundamental opportunities and challenges posed by 3-D ICs especially from the context of design tools and methods. We also discuss the methodology of co-design to address more complicated and interdependent design problems in 3-D IC, and conclude with a discussion of the remaining challenges and open problems that must be overcome to make 3-D IC technology commercially viable.

Monolithic 3-D (M3D) integrated circuits (ICs) are an emerging technology that offer much higher integration densities than previous 3-D IC approaches. In this paper, we present a complete netlist-to-layout design flow to design an M3D block, as well as to integrate 2-D and 3-D blocks into an M3D SoC. This design flow is based on commercial tools built for 2-D ICs, and enhanced with our 3-D specific methodologies. We use the OpenSPARC T2 SoC as a case study, implement it in a 28-nm fully depleted silicon on insulator foundry process, and demonstrate that we can achieve up to 12% and 8% power savings for a single block and SoC, respectively, when compared with their 2-D counterparts implemented using commercial tools.

Monolithic 3-D ICs (M3-D ICs) offer extremely high vertical interconnection density, significantly improving the power-performance envelope when compared to conventional 2-D ICs. However, process limitations lead to one tier having either degraded transistors or interconnects. This paper models the amount of degradation that can be expected at current and future nodes (45, 22, and 10 nm), develops a process development kit using these models to enable evaluation, and presents a block-level M3-D IC RTL-to-GDSII flow that is capable of mitigating some of this degradation. Experimental results indicate that at lower technology nodes, M3-D ICs offer more benefits. Results also indicate that the impact of transistor degradation is diminished at lower technology nodes while the impact of interconnect degradation becomes worse. Overall, perfect M3-D ICs close more than half the gap in the power-performance envelope between 2-D ICs and the “ideal” block-level design. While degraded tiers reduce the benefit of M3-D ICs, our degradation-aware floorplanner gives back up to 17% of the loss, and helps to obtain significant overall benefits compared to 2-D ICs.

In this paper, we present a comprehensive study on the impact of power delivery network (PDN) on full-chip wirelength, routability, power, and thermal effects in gate-level monolithic 3-D (M3-D) ICs across different technology nodes. Our studies show that PDN worsens routing congestion more severely in M3-D ICs than in 2-D designs due to the significant reduction in resources for 3-D connections. The relative impact worsens at advanced technology nodes due to higher congestion of interconnects. The increase in signal wirelength translates into additional net switching power dissipation, which significantly contributes to total power. This in turn aggravates thermal issues further in 3-D ICs. In addition, we observe that PDN tradeoffs among wirelength, power, and thermal are more pronounced in M3-D ICs than through silicon via-based 3-D and 2-D designs because of the higher integration density and the severe competition between signal and power connections. We also compare the impact of PDN on full chip routing in M3-D ICs versus face-to-face 3-D ICs. Lastly, we use various PDN design optimization techniques for M3-D ICs at different nodes and obtain up to 13.9% signal wirelength and 17.6% total power reduction under the given IR drop budget.

Monolithic 3-D (M3D) IC is one of the potential technologies to break through the challenges of continued circuit power and performance scaling. In this paper, for the first time, we demonstrate the power benefits of M3D and present design guideline in a 7-nm FinFET technology node. The predictive 7-nm process design kit (PDK) and the standard cell library using both high-performance (HP) and low-standby-power (LSTP) device technologies are developed based on NanGate 45-nm PDK using accurate dimensional, material, and electrical parameters from publications and a commercial-grade tool flow. We implement full-chip M3D designs utilizing industry-standard physical design tools, and gauge the impact of M3D technology on performance, power, and area metrics. We also provide the design guidelines as well as a new partitioning methodology to improve M3D design quality. This paper shows that M3D designs outperform 2-D counterparts by 16% and 16.5% on average in terms of isoperformance total power reduction with 7-nm HP and LSTP cell library, respectively. This demonstrates the power benefits of M3D technology in both HP and low-power future generation devices.

A chain-based coupling delay estimation method for through-silicon-vias (TSVs) in 3-D integrated circuits is proposed. Existing works target the worst case scenarios and this leads to inaccurate TSV coupling delay estimations, as the worst case may not occur during normal operation. The proposed method calculates the TSV coupling delay using simulation-based switching data. In addition, our TSV chain method allows us to capture the effects of nonneighboring TSVs accurately. Our simulations show that the error introduced by our method without using HSPICE is less than 10 ps even in TSV-crowded regions.

Low power is considered as the driving force for 3-D ICs, yet there have been few thorough design studies on how to reduce power in 3-D ICs. In this paper, we discuss computer-aided design techniques and design methodologies to reduce power consumption in 3-D IC designs using a commercial grade CPU core (OpenSPARC T2 core). To demonstrate power benefits in 3-D ICs, four design techniques are explored: 1) 3-D floorplanning; 2) metal layer usage control for intrablock-level routing; 3) dual-Vth design; and 4) functional unit block (FUB) folding. The benefits and challenges of multiple FUB folding are also discussed. Finally, the through-silicon via technology scaling impact on FUB folding and 3-D power benefit is examined. With the aforementioned methods combined, our 2-tier 3-D designs provide up to 52.3% reduced footprint, 27.9% shorter wirelength, 35.4% decreased buffer cell count, and 27.8% power reduction over the 2-D counterpart under the same performance.

Monolithic 3D IC (M3D) has the potential to provide a break-through in the power and performance scaling challenges. We, for the first time, present a comprehensive study of M3D on a commercial design across multiple technology nodes. The performance and power impact of M3D is investigated using a commercial, in-order, 32-bit application processor, implemented on foundry 28nm and 14/16nm process nodes, as well as a predictive 7nm node. We study the factors across the technology nodes that affect the efficiency of M3D, and propose a roadmap for optimum technology and design interaction that will enable the full entitlement of M3D.

Monolithic 3D IC (M3D) can continue to improve power, performance, area and cost beyond traditional Moore's law scaling limitations by leveraging the third-dimension and fine-grained monolithic inter-tier vias (MIVs). Several recent studies present methodologies to implement M3D designs, but most, if not all of these studies implement top and bottom tier separately after partitioning, which results in inaccurate buffer insertion. In this paper, we present a new methodology called ‘Cascade2D’ that utilizes design and micro-architecture insight to partition and implement an M3D design using 2D commercial tools. By modeling MIVs with sets of anchor cells and dummy wires, we implement and optimize both top and bottom tier simultaneously in a single 2D design. M3D designs of a commercial, in-order, 32-bit application processor at the foundry 28nm, 14/16nm and predictive 7nm technology nodes are implemented using this new methodology and we investigate the power, performance and area improvements over 2D designs. Our new methodology consistently outperforms the state-of-the-art M3D design flow with up to 4× better power savings. In the best case scenario, M3D designs from the Cascade2D flow show 25% better performance at iso-power and 20% lower power at isoperformance.

In this paper we study power, performance, and cost (PPC) tradeoffs for 2-tier, gate-level, full-chip GDS monolithic 3D ICs (M3D) built using a foundry-grade 7nm bulk FinFET technology. We first develop highly-accurate wafer and die cost models for 2D and M3D to study PPC tradeoffs. In our study, both 2D and M3D designs are optimized in terms of the number of BEOL metal layers used for routing to obtain the best possible PPC values. We develop a new CAD methodology for 2-tier gate-level M3D, named Projected 2D Flow, that allows us to accurately compare RC parasitics of equivalent nets in both 2D and M3D designs. Our experiments based on two different circuit types (BEOL-dominant vs. FEOL-dominant) confirm that M3D designs indeed offer a significant footprint saving. However, to our surprise, the PPC quality of M3D turns out to be worse than that of 2D by 34% due to the high wafer cost of M3D. Our study also reveals that M3D wafer yield should be as high as 90% of 2D wafer yield, and the M3D device manufacturing cost should be less than 33% of that of 2D to justify the adoption of M3D technology at the 7nm era. Lastly, and counter-intuitively, our study shows that FEOL-dominant circuit shows more PPC benefits from M3D technology than BEOL-dominant circuit.

In this paper, we develop tier partitioning strategy to mitigate back-end-of-line (BEOL) interconnect delay degradation and cost issues in monolithic 3D ICs (M3D). First, we study the routing overhead and delay degradation caused by tungsten BEOL interconnect in the bottom-tier of M3D. Our study shows that tungsten BEOL reduces performance by up to 30% at 4× resistance increase of bottom-tier interconnect. In addition, the bottom-tier BEOL adds a routing overhead to 3D nets, which is neglected in the state-of-the-art flow. Next, we develop two partitioning methods targeted specifically towards BEOL impact reduction. Our path-based approach identifies critical timing paths and places their cells in the top-tier to reduce the impact of delay degradation and routing overhead. Our net-based partitioning methodology confines the nets with long 2D wirelength into the top-tier to reduce the overall routing demand, and hence the metal layer usage in the bottom-tier. This in turn results in BEOL cost savings. Using a foundry 22nm FDSOI technology and full-chip GDS designs, we achieve tolerance of up to 4× increase in the bottom-tier BEOL resistance using our partitioning strategy. In addition, we save up to 3 metal layers in the bottom-tier of our M3D designs with up to 32% power savings over 2D IC for an interconnect dominated benchmark.

Monolithic 3D IC is an emerging technology to continuously satisfy demands for power reduction under challenges posed by traditional device scaling. In this paper, for the first time, we study power benefits of 4-tier monolithic 3D ICs compared with 2-tier monolithic 3D and 2D ICs. We present a tier partitioning methodology that significantly extends the capability of a state-of-the-art flow built for 2-tier monolithic 3D ICs. We develop two complete RTL-to-GDSII design flows to achieve this goal and offer quantitative comparisons. In addition, we study impacts of inter-tier via usage on 2-tier and 4-tier monolithic 3D ICs. Our experiments show that poorly controlled inter-tier via usage results in up to 6.05% degradation in total power savings. Thus, we develop an effective strategy to achieve inter-tier via configurations to optimize power metrics. Experiments show that 4-tier monolithic 3D ICs outperform 2-tier and 2D IC by 15% and 50% in terms of power and 25% and 75% in area under the same performance.

In this paper, we develop physical design tools and methodologies to tackle the inter-tier performance variations caused by low temperature manufacturing in 2-tier gate-level monolithic 3D ICs (M3D). First, we model the top tier front-end-of-line (FEOL) device mobility degradation and its impact on cell delay/power values. Next, we quantify the impact of tungsten interconnect and cost-driven metal layer saving in the back-end-of-line (BEOL) of the bottom tier. These device and interconnect degradation models are used in our new full-chip M3D physical design flow named Derated 2D. This flow overcomes the well-known drawback of the state-of-the-art Shrunk 2D that requires shrinking of layout objects and RC parasitics. Also, Derated 2D performs low-temperature process-aware tier partitioning to effectively keep timing-critical components in the bottom tier. Moreover, Derated 2D conducts timing-driven monolithic inter-tier via (MIV) planning to cope with the resistivity increase in tungsten BEOL. Lastly, Derated 2D offers an effective timing closure solution through a post-route optimization. Experiments based on a foundry-grade 7nm FinFET process design kit (PDK) show that Derated 2D achieves up to 36% performance improvement and 10% energy saving compared with Shrunk 2D. Using a post-route optimization, Derated 2D further improves timing under various FEOL/BEOL degradation settings at a minimum energy overhead.

In this paper we study the impact of low thermal budget process on design quality in monolithic 3D ICs (M3D). Specifically, we quantify how much the tier-to-tier transistor performance difference affects full-chip power and performance metrics in a foundry 14nm FinFET technology. Our study first shows that 5%, 10%, and 15% top-tier device degradation in a wire-dominated, timing-closed monolithic 3D IC design leads to 7%, 12%, and 18% full-chip timing violation, respectively. Next, we address this impact with our CAD solution named Tier-Aware M3D (TA-M3D) flow that identifies potential timing-critical paths and partitions them into the faster (bottom) tier to minimize the top-tier degradation impact. One unique challenge in timing closure in this case, is how to conduct buffering and sizing on the paths that lie entirely in the top or bottom-tier as well as those that span both tiers. Our approach handles all 3 types of paths carefully and closes timing under the given top-tier degradation assumption, while minimizing the total power consumption. Our enhanced monolithic 3D IC designs, even with 5%, 10%, and 15% slower transistors in the top-tier, still offers 26%, 24%, and 5% power savings over 2D IC, respectively. Our study also covers other types of circuits.

In this paper, we first present a comprehensive study of the unique thermal behavior in monolithic 3-D integrated circuits (ICs) in contrast to through silicon via-based 3-D ICs. In particular, we study the impact of the thin interlayer dielectric between the device tiers on vertical thermal coupling. We then study and compare the impact of different application-based package structures on the thermal behavior of monolithic 3-D ICs. With these unique properties and behavior, we develop a fast and accurate compact full-chip thermal analysis model based on nonlinear regression technique which adapts to the package structure during development and hence considers it during temperature evaluation. Our model is extremely fast and highly accurate with an error of less than 5%. This model is incorporated into a thermal-aware 3-D-floorplanner that runs without significant runtime overhead. We use the floorplanner with our package-aware thermal model and observe up to 22% reduction in the maximum temperature with insignificant area and performance overhead.

Low-power is one of the key driving forces in modern very large scale integration systems. Recent studies show that 3-D integrated circuits (ICs) offer a significant power saving over 2-D ICs. However, these studies are mainly limited to two-tier (2-tier) designs. Thus, in this paper, we extend our target to three-tier (3-tier) 3-D ICs. This paper first shows that the one additional tier available in 3-tier 3-D ICs does offer more power saving compared with their 2-tier 3-D IC counterparts, but more careful floorplanning, through-silicon via management, and block folding considerations are required. Second, we find that the 3-tiers can be bonded in several different ways: 1) face-to-back only; 2) face-to-face and face-to-back combined; and 3) back-to-back and face-to-face combined. This paper shows that these choices pose additional challenges in design optimizations for more power saving. Lastly, we develop effective computer-aided-design solutions that are seamlessly integrated into commercial 2-D IC tools to handle 3-tier 3-D IC power optimization under various bonding style options. With our low-power design methods combined, our 3-tier 3-D ICs provide -14.8% more power reduction over 2-tier 3-D ICs, and -36.0% over 2-D ICs in microprocessor cores under the same performance. In full-chip microprocessors, our 3-tier 3-D ICs provide -27.2% more power reduction over 2-D ICs.

Through-silicon-via (TSV)-to-TSV coupling is a new phenomenon in 3-D ICs, and it becomes a significant source of signal integrity problems. The existing studies on its extraction and analysis, however, become inaccurate when handling more than two TSVs on full-chip scale. In this paper, we investigate the multiple TSV-to-TSV coupling issue and propose a model that can be efficiently used for full-chip extraction. Then, we perform an analysis on the impact of TSV parasitics on coupling and delay. Unlike the common belief that only the closest neighboring TSVs affect the victim, this paper shows that nonneighboring aggressors also cause nonnegligible coupling noise. Based on these observations, we propose an effective method of reducing the overall coupling level.

Power delivery to the tier farthest away from the package in 3D VLSI is challenging. This is because the current provided by the package on the bottom is (1) first used by other tiers before it reaches the top, and (2) delivered using extremely small-size intra and inter-tier vias. Our solution is a tier partitioning method that assigns power hungry cells to the tier closer to the package, which is farther away from the heat spreader. Our study shows that this approach alleviates the IR-drop, power delivery network (PDN) resource usage, and power consumption in the top tier. Moreover, moving the cells to the bottom tier, unlike popular belief, does not cause any serious thermal issues. This is especially true in mobile applications, where heat is dissipated by both the heat spreader and printed circuit board. In summary, our tier-partitioning leads to 24.66% IR-drop reduction, 28.57% PDN resource reduction, and 4% wirelength reduction, with < 1°C increase in temperature.

3D DRAM is the next-generation memory system targeting high bandwidth, low power, and small form factor. This paper presents a cross-domain CAD/architectural platform that addresses DC power noise issues in 3D DRAM targeting stacked DDR3, Wide I/O, and hybrid memory cube technologies. Our design and analysis include both individual DRAM dies and a host logic die that communicates with them in the same stack. Moreover, our comprehensive solutions encompass all major factors in design, packaging, and architecture domains, including power delivery network wire sizing, redistribution layer routing, distributed, and dedicated TSV placement, die bonding style, backside wire bonding, and read policy optimization. We conduct regression analysis and optimization to obtain high quality solutions under noise, cost, and performance tradeoff. Compared with industry standard baseline designs and policies, our methods achieve up to 68.2% IR-drop reduction and 30.6% performance enhancement.

Face-to-face (F2F) bonded 3D ICs are promising design solutions. However, because of the short die-to-die distance, direct coupling between the metal layers of the top and bottom dies introduces severe signal integrity problems that call for accurate extraction. This study is the first to demonstrate and compare three parasitic extraction methods of F2F-bonded 3D ICs. One is traditional die-by-die extraction, which cannot handle inter-die coupling and E-field sharing. We propose another method, holistic extraction, which treats all layers from both dies simultaneously and captures all inter-die coupling at the cost of high Layout Versus Schematic (LVS) complexity. We also propose an in-context extraction method that accounts for interface layers between dies. Carefully handling double-counting and surface layers issues, in-context extraction is LVS-friendly without losing accuracy. Full-chip analyses show that both of our extraction methods are highly accurate and able to handle various metal layers in several process nodes. It also corrects timing, power, and signal integrity errors introduced by die-by-die extraction. In-context extraction with two interface layers is highly accurate and efficient with an error of 0.9% for total ground capacitance and 0.8% for total coupling capacitance.

Low-power is one of the key driving forces in modern VLSI systems. Several recent studies show that 3D ICs offer significant power savings over 2D ICs, primarily due to wirelength and buffer saving. However, these existing studies are mainly limited to 2-tier designs. In this paper, our target is extended to 3-tier 3D ICs. Our study first shows that the one additional tier available in 3-tier 3D ICs does offer more power saving compared with their 2-tier 3D IC counterparts, but more careful floorplanning, through-silicon via (TSV) management, and block folding considerations are required. Second, we find that the three tiers can be bonded in different ways: (1) face-to-back only and (2) face-to-face and face-to-back combined. Our study shows that these choices pose additional challenges in design optimizations for more power saving. Lastly, we develop effective CAD solutions that are seamlessly integrated into commercial 2D IC tools to handle 3-tier 3D IC power optimization under various bonding style options. With our low-power design methods combined, our 3-tier 3D ICs provide -14.8% more power reduction over 2-tier 3D ICs and -36.0% over 2D ICs under the same performance.

Through-silicon vias (TSVs) are subject to thermal fatigue due to stress over time, no matter how small the stress is. Existing works on TSV fatigue all rely on measurement-based parameters to estimate the lifetime, and cannot consider detailed thermal profiles. In this paper, we propose a new method for TSV fatigue prediction using entropy production during thermal cycles, which is validated by theoretical analysis and measurement results. By combining thermodynamics and mechanics laws, the fatigue process can be quantitatively evaluated with detailed thermal profiles. Experimental results show that interestingly, the landing pad possesses the most easy-to-fail region, which generates up to 50% more entropy compared with the TSV body. The impact of landing pad dimension and TSV geometries are also studied, providing guidance for reliability enhancement. Finally, full-chip fatigue analysis is performed based on stress superposition. To the best of the authors' knowledge, this is the first TSV fatigue model that is free of measurement data fitting, the first that is capable of considering detailed thermal profiles, and the first framework for efficient full-chip TSV fatigue analysis.

In three-dimensional integrated circuit (3D IC) systems that use through-silicon via (TSV) technology, a significant design consideration is the coupling noise to/from TSVs. In Analog/Mixed signal ICs, the TSV coupling effect can cause coupling noise disturbance and degrades the performance of sensitive analog devices. In this paper, two different stacking integrations for the sensor array and the 3D IC, are considered and compared. In the face-up stacking integration, the transducers' flip chip bonded pads are directly bonded to the integrated circuit. In facedown integration type, however, the connections from the sensor output signal to the front end electronics is done via the TSVs. The TSV coupling noise is compared for the two schemes using the existing TSV coupling noise model. To validate the impact of stacking integration on the coupling noise, an ADC case study was designed and implemented using 130nm device technology and Tezzaron TSV technology. The simulations results show that the face-up integration could suppress the coupling noise by 10db. Moreover, a 23% reduction in footprint is achieved in this stacking integration.

Monolithic 3D IC (M3D) is one potential technology to help with the challenges of continued circuit power and performance scaling. In this paper, for the first time, the power benefits of monolithic 3D IC (M3D) using a 7nm FinFET technology are investigated. The predictive 7nm Process Design Kit (PDK) and standard cell library for both high performance (HP) and low standby power (LSTP) device technologies are built based on NanGate 45nm PDK using accurate dimensional, material, and electrical parameters from publications and a commercial-grade tool flow. In addition, we implement full-chip M3D GDS layouts using both 7nm HP and LSTP cells and industry-standard physical design tools, and evaluate the resulting full-chip power, performance, and area metrics. Our study first shows that 7nm HP M3D designs outperform 7nm HP 2D designs by 16.8% in terms of iso-performance total power reduction. Moreover, 7nm LSTP M3D designs reduce the total power consumption by 14.3% compared to their 2D counterparts. This convincingly demonstrates the power benefits of M3D technologies in both high performance as well as low power future generation devices.

In three-dimensional (3D) integrated circuits (IC-s), many clock-TSVs are deployed to deliver clock signals to different tiers with minimum skews. However, these clock-TSVs are prone to aging effects, such as thermal-mechanical stress and electromigration, rendering hard-to-predict clock skews at runtime. These skews have a wide range of influence on the flip-flops, and may violate the safety margins of critical paths in the circuit. Besides the circuit aging effect, the clock-TSV induced skews pose another threat to the circuit lifetime reliability. To tackle this problem, we propose to put tunable buffer for each clock-TSV in the clock network, and introduce an efficient algorithm to place aging sensors in the circuit at design stage. Then, at runtime, we conduct online diagnosis and apply effective clock tuning algorithms based on the triggered alarms in the aging sensors. Experimental results on a post-layout 3D circuit show that the proposed solution is able to significantly improve the lifetime reliability of 3D ICs.

This paper describes the architecture, design, analysis, and simulation and measurement results of the 3D-MAPS (3D massively parallel processor with stacked memory) chip built with a 1.5 V, 130 nm process technology and a two-tier 3D stacking technology using 1.2$\micro\hbox{m}$-diameter, 6$\micro \hbox{m}$-height through-silicon vias (TSVs) and$3.4\nbsp\micro\hbox{m}$-diameter face-to-face bond pads. 3D-MAPS consists of a core tier containing 64 cores and a memory tier containing 64 memory blocks. Each core communicates with its dedicated 4KB SRAM block using face-to-face bond pads, which provide negligible data transfer delay between the core and the memory tiers. The maximum operating frequency is 277 MHz and the maximum memory bandwidth is 70.9 GB/s at 277 MHz. The peak measured memory bandwidth usage is 63.8 GB/s and the peak measured power is approximately 4 W based on eight parallel benchmarks.

Monolithic 3D (M3D) is an emerging technology that enables integration density which is orders of magnitude higher than that offered by through-silicon-vias. In this paper, we demonstrate that a modified 2D placement technique coupled with a post-placement partitioning step is sufficient to produce high-quality M3D placement solutions. We also present a commercial router-based monolithic intertier via insertion methodology that improves the routability of M3D ICs. We demonstrate that, unlike in 2D ICs, the routing supply and demand in M3D ICs are not completely independent of each other. We develop a routing demand model for M3D ICs, and use it to develop an O(N) min-overflow partitioner that enhances routability by off-loading demand from one tier to another. This technique reduces the routed wirelength and the power delay product by up to 7.44% and 4.31%, respectively. This allows a two-tier M3D IC to achieve, on average, 19.9% and 11.8% improvement in routed wirelength and power delay product over 2D, even with reduced metal layer usage.

The through-silicon-via (TSV) introduces new parasitic components into 3-D ICs. This paper presents a novel method of extracting the parasitic capacitance between TSVs and their surrounding wires. For the first time, we examine electrical field (E-field) sharing effects from multiple TSVs and neighboring wires and their impact on timing, power, and noise with full-chip sign-off analyses. For fast and accurate full-chip extraction, we propose a pattern-matching algorithm that accounts for the physical dimensions of multiple TSVs and neighboring wires and captures all E-field interactions. Compared with the average error of a field solver, that of our extraction method, which requires only 2.4 s runtime and negligible memory for a full-chip 64-point fast Fourier transform (FFT64) design with 330 TSVs, is 0.063fF. Upon extraction of TSV-related parasitics, we observe that TSV-to-wire capacitance significantly increase average TSV net noise and the longest path delay. To reduce TSV-to-wire coupling, we implement two full-chip optimization methods and show that increasing the minimum distance between TSVs and neighboring wires reduces both coupling noise and the aggressor count. Thanks to E-field sharing from grounded wire guard rings, victim TSVs are more effectively shielded from aggressor noise. A full-chip analysis shows that these methods are highly effective in reducing noise with only slight impact on timing and area.

Prebond testing of through-silicon-vias (TSVs) and die logic is a significant challenge and a potential roadblock for 3-D integration. Built-in self-test solutions introduce considerable die area overhead. Oversized probe pads on TSVs to provide prebond test access limit both test bandwidth and TSV density. This paper presents a solution to these problems, allowing a probe card to contact TSVs without the need for probe pads, enabling both TSV and prebond scan test. Several possible prebond scan test configurations are shown-they provide varying degrees of test parallelism under design constraints. HSPICE simulations are performed on two logic-on-logic 3-D benchmarks. Results show that the ratio of the number of probe needles available for test access to the number of prebond scan chains determines which prebond scan configuration results in the shortest test time. Maximum prebond scan-in and scan-out shift-clock speeds are determined for dies in a benchmark 3-D design as a function of driver strength and transmission gate width. These clock speeds show that prebond scan test can be performed at a speed that is comparable with scan testing of packaged dies. The maximum clock speed can also be tuned by changing the drive strength of the probe and on-die drivers of the TSV network. Estimates are also provided for peak and average power consumption during prebond scan test for both a high-power pattern per scan chain and an average power pattern per scan chain. On-die area overhead for the proposed method is estimated to be between 1.0% and 2.9% per die for two 3-D benchmarks.

Low power is widely considered as a key benefit of 3D ICs, yet there have been few thorough design studies on how to maximize power benefits in 3D ICs. In this paper, we present design methodologies to reduce power consumption in 3D ICs using a large-scale commercial-grade microprocessor (OpenSPARC T2). To further improve power benefits in 3D ICs on top of the traditional 3D floorplanning, we study the impact of block folding and bonding styles. We also develop an effective method to place face-to-face vias for our 2-tier 3D design for power optimization. With aforementioned methods combined, our 3D designs provide up to 20.3% power reduction over the 2D counterpart under the same performance.

In this paper we study the power vs. performance tradeoff in block-level monolithic 3D IC designs. Our study shows that we can close the power-performance gap between 2D and a theoretical lower bound by up to 50%. We model the inter-tier performance variations caused by a low temperature manufacturing process on the non-bottom tiers. We also model an alternate manufacturing process, where highly resistive tungsten interconnects are used on the bottom tier to withstand a high temperature process on the non-bottom tiers. We propose a variation-aware floorplanning technique that makes our design more tolerant to these variations. We demonstrate that our design methods can help us obtain high quality designs even under inter-tier performance variations.

In this paper, for the first time, we model and extract the parasitic capacitance between TSVs and their surrounding wires in 3D IC. For a fast and accurate full-chip extraction, we propose a pattern-matching-based algorithm that considers the physical dimensions of TSVs and neighboring wires and captures their field interactions. Our extraction method is accurate within 1.9% average error for a full-chip-level design while requiring negligible runtime and memory compared with a field solver. We also observe that TSV-to-wire capacitance has a significant impact on the noise of TSV-based connections and the longest path delay. To reduce TSV-to-wire coupling, we present two full-chip optimization methods, i.e., increasing KOZ and guard ring protection that are shown to be highly effective in noise reduction with minimal overhead.

In this paper, we present a comprehensive study of the unique thermal behavior in monolithic 3D ICs. In particular, we study the impact of the thin inter-layer dielectric (ILD) between the device tiers on vertical thermal coupling. In addition, we develop a fast and accurate compact full-chip thermal analysis model based on non-linear regression technique. Our model is extremely fast and highly accurate with an error of less than 5%. This model is incorporated into a thermal-aware 3D-floorplanner that runs without significant runtime overhead. We observe up to 22% reduction in the maximum temperature with insignificant area and performance overhead.

The complex buses consume significant power in graphics processing units (GPUs). In this paper, we demonstrate how the power consumption of buses in GPUs can be reduced with 3D IC technologies. Based on layout simulations, we found that partitioning and floorplanning of 3D ICs affect the power benefit amount, as well as the technology setup, target clock frequency, and circuit switching activity. For 3D IC technologies using two dies, we achieved the total power reductions of up to 21.5% over a baseline 2D design.

In this paper, we present a comprehensive study on the impact of power delivery network (PDN) on full-chip wirelength, routability, power, and thermal effects in monolithic 3D ICs. Our studies first show that the full PDN worsens routing congestion more severely in monolithic 3D ICs than in 2D designs due to the significant reduction in resources for 3D connections. The increase in signal wirelength translates into additional net switching power dissipation, which significantly contributes to total power. This in turn aggravates thermal issues in 3D ICs. In addition, we observe that PDN tradeoffs among wirelength, power, and thermal are more pronounced in monolithic 3D ICs than TSV-based 3D and 2D designs. This is because of the higher integration density and the severe competition between signal and power connections. Lastly, we develop various PDN design optimization techniques for monolithic 3D ICs and obtain up to 8% signal wirelength and 5% maximum temperature reduction under the given IR drop budget.

In a gate-level monolithic 3D IC (M3D), all the transistors in a single logic gate occupy the same tier, and gates in different tiers are connected using nano-scale monolithic inter-tier vias. This design style has the benefit of the superior power-performance quality offered by flat implementations (unlike block-level M3D), and zero total silicon area overhead compared to 2D (unlike transistor-level M3D). In this paper we develop, for the first time, a complete RTL-to-GDSII design flow for gate-level M3D. Our tool flow is based on commercial tools built for 2D ICs and enhanced with our 3D specific methodologies. We use this flow along with a 28nm PDK to build layouts for the OpenSPARC T2 core. Our simulations show that at the same performance, gate-level M3D offers 16% total power reduction with 0% area overhead compared to commercial quality 2D IC designs.

Monolithic 3D is an emerging technology that enables integration density which is orders of magnitude higher than that offered by through-silicon-vias (TSV). In this paper we demonstrate that a modified 2D placement technique, coupled with a post-placement partitioning step, is sufficient to produce high quality monolithic 3D placement solutions. We also present a commercial router based monolithic inter-tier via (MIV) insertion methodology that dramatically improves the routability of monolithic 3D-ICs. We develop a routing demand model for monolithic 3D-ICs, and use it to develop an O(N) min-overflow partitioner that enhances routability by off-loading demand from one tier to another. This technique reduces the routed wirelength and the power delay product (PDP) by up to 4% and 4.33% respectively, under the same half-perimeter wirelength. This allows a two-tier monolithic 3D-IC to achieve, on average, 19.2% and 12.1% improvement in routed wirelength and PDP over 2D, even with reduced metal layer usage.

3-D integrated circuits (3-D ICs) are expected to have shorter wirelength, better performance, and less power consumption than 2-D ICs. These benefits come from die stacking and use of through-silicon vias (TSVs) fabricated for interconnections across dies. However, the use of TSVs has several negative impacts such as area and capacitance overhead. To predict the quality of 3-D ICs more accurately, TSV-aware 3-D wirelength distribution models considering the negative impacts were developed. In this paper, we apply an optimal buffer insertion algorithm to the TSV-aware 3-D wirelength distribution models and present various prediction results on wirelength, delay, and power consumption of 3-D ICs. We also apply the framework to 2-D and 3-D ICs built with various combinations of process and TSV technologies and predict the quality of today and future 3-D ICs.

Electromigration (EM) in power distribution networks (PDNs) is a major reliability issue in 3-D ICs. While the EM issues of local vias and through-silicon-vias (TSV) have been studied separately, the interplay of TSVs and conventional local vias in 3-D ICs has not been well investigated. This co-design is necessary when the die-to-die vertical power delivery is done using both TSVs and local interconnects. In this paper, we model EM for PDNs of 3-D ICs with a focus on multiscale via (MSV) structure, i.e., TSVs and local vias used together for vertical power delivery. We study the impact of structure, material, and preexisting void conditions on the EM-related lifetime of our MSV structures. We also investigate the transient IR-voltage change of full-chip level 3-D PDNs with MSVs with our model. The experimental results demonstrate that our EM modeling can effectively capture the EM reliability of the full-chip level 3-D PDNs with MSVs, which can be hard to achieve by the traditional EM analysis based on the individual local via or the TSV.

This paper presents a silicon effect-aware multiTSV model. Through-silicon-via (TSV) depletion region, silicon substrate discharging path and electrical field distribution around TSV neighbor are modeled and studied in full-chip design. Verification with field solver and full-chip TSV-to-TSV coupling analysis in both the worst case and the average case show this model is accurate and efficient. It is found that 3-D nets receive more noise than their 2-D counterparts due to TSV-to-TSV coupling. To alleviate this coupling noise on TSV nets, two new optimization methods are investigated. One way is to utilize guard rings around the victim TSV so as to form a stronger discharging path, an alternative approach is to adopt differential signal transmission to improve noise immunity. These techniques have been implemented on 3-D IC designs with TSVs placed regularly or irregularly. Full-chip analysis results show that our approaches are effective in noise reduction with small area overhead.

In this paper, we propose two methods used in 3-D IC placement that efficiently exploit the die-to-die thermal coupling in the stack. First, through-silicon vias (TSVs) are spread on each die to reduce the local power density and vertically aligned across dies simultaneously to increase thermal conductivity to the heatsink. Second, we move high-power logic cells to the location that has higher conductivity to the heatsink while moving TSVs in the upper dies so that high-power cells are vertically overlapping below the TSVs. These methods are employed in a force-directed 3-D placement successfully and outperform several state-of-the-art placers published in recent literature. We obtain 3-D placement results with shorter routed wirelength at similar temperature. We also obtain 3-D placement results with lower temperatures at similar routed wirelengths.

Backend dielectric breakdown degrades the reliability of circuits. A methodology to estimate chip lifetime because of backend dielectric breakdown is presented. It incorporates failures because of parallel tracks, the width effect, and field enhancement due to line ends. It also includes the operating temperature and activity.

In 3D ICs, block-level designs provide various advantages over designs done at other granularity such as gate-level because they promote the reuse of IP blocks. In this paper, we study block-level 3D-IC designs, where the footprint of the dies in the stack are different. This happens in case of die-to-wafer bonding, which is more popular choice for near-term low-cost 3D designs. We study design quality tradeoffs among three different ways to place through-silicon vias (TSVs): TSV-farm, TSV-distributed, and TSV-whitespace. In our holistic approach, we use wirelength, power, performance, temperature, and mechanical stress metrics to conduct comprehensive comparative studies on the three design styles. In addition, we provide analysis on the impact of TSV size and pitch on the design quality of these three styles.

Three dimensional integrated circuits (3D-ICs) have emerged as a promising solution to continue device scaling. They can be realized using Through Silicon Vias (TSVs), or monolithic integration using Monolithic Inter-tier vias (MIVs), an emerging alternative that provides much higher via densities. In this paper, we provide a framework for floorplanning existing 2D IP blocks into 3D-ICs using MIVs. We take the floorplanning solution all the way through place-and-route and report post-layout metrics for area, wirelength, timing, and power consumption. Results show that the wirelength of TSV-based 3D designs outperform 2D designs by upto 14% in large-scale circuits only. MIV-based 3D designs, however, offer an average wirelength improvement of 33% for a wide range of benchmark circuits. We also show that while TSV-based 3D cannot improve the performance and power unless the TSV capacitance is reduced, MIV-based 3D offers significant reduction of upto 33% in the longest path delay and 35% in the inter-block net power.

3D physical design needs accurate device model of through-silicon vias (TSVs). In this paper, physics-based electrical-thermal model is introduced for both signal and dummy thermal TSVs with the consideration of nonlinear electrical-thermal dependence. Taking thermal-reliable 3D clock-tree synthesis as a case-study to verify the effectiveness of the proposed TSV model, one nonlinear programming-based clock-skew reduction problem is formulated to allocate thermal TSVs for clock-skew reduction under non-uniform temperature distribution. With a number of 3D clock-tree benchmarks, experiments show that under the nonlinear electrical-thermal TSV model, insertion of thermal TSVs can effectively reduce temperature-gradient introduced clock-skew by 58.4% on average, and has 11.6% higher clock-skew reduction than the result under linear electrical-thermal model.

The nano-scale 3D interconnects available in monolithic 3D IC technology enable ultra-high density device integration at the individual transistor-level. In this paper we demonstrate the power benefits of transistor-level monolithic 3D designs. We first build a cell library that consists of 3D gates and model their timing/power characteristics. Next, we build timing-closed, full-chip GDSII layouts and perform sign-off iso-performance power comparisons with 2D IC designs. We also study the characteristics of benchmark circuits that maximize the power benefits in monolithic 3D designs. Lastly, our study is extended to predict the power benefits of monolithic 3D designs built with future devices.

TSV-to-TSV coupling is a new parasitic element in 3D ICs and can become a significant source of signal integrity problem. Existing studies on its extraction, however, becomes highly inaccurate when handling more than two TSVs on full-chip scale. In this paper we investigate the multiple TSV-to-TSV coupling issue and propose an accurate model that can be efficiently used for full-chip extraction. Unlike the common belief that only the closest neighboring TSVs affect the victim, our study shows that non-neighboring aggressors also cause non-negligible impact. Based on this observation, we propose an effective method of reducing the overall coupling level in multiple TSV cases.

Electromigration (EM) in power distribution network (PDN) is a major reliability issue in 3D ICs. While the EM issues of local vias and through-silicon-vias (TSV) have been studied separately, the interplay of TSVs and conventional local vias in 3D ICs has not been well investigated. This co-design is necessary when the die-to-die vertical power delivery is done using both TSVs and local interconnects. In this work, we model EM for PDN of 3D ICs with a focus on multi-scale via structure, i.e., TSVs and local vias used together for vertical power delivery. We study the impact of structure, material, and pre-existing void conditions on EM-related lifetime of our multi-scale via structures. Experimental results demonstrate that our EM modeling can effectively capture the EM reliability of the entire multi-scale via in 3D PDN, which can be hard to achieve by the traditional EM analysis based on the individual local via or TSV.

In this paper, we present a multiple-TSV based TSV-to-TSV coupling model and extraction methods that consider the impact of depletion region, the silicon substrate effect, and the electrical field distribution around TSVs. Our studies show that these factors have a significant impact on the individual and full-chip scale TSV-to-TSV coupling. Our effort leads to a simplified coupling model that is accurate and efficient on timing, power, and signal integrity in full-chip scale. In order to alleviate the coupling noise in full-chip level 3DIC, we propose grounded guard rings that are more effective than grounded TSV insertion. Results show that our approach reduces coupling noise on TSV nets up to 27.3% with only 7.65% area overhead.

The CTE mismatch-induced stress in 3D ICs may initiate cracks from the interface between a TSV and its dielectric liner, and propagates them on the silicon substrate surface. If a crack grows beyond the keep-out-zone (KOZ) of a TSV, it will jeopardize the reliability of the devices along its propagation path. While such threat can be eliminated by a sufficiently large KOZ, significant area overhead will be incurred. Given the low probability of crack occurrence, we argue that a much more economical approach is to keep KOZ small and filter out bad chips with cracks growing beyond the KOZ during testing. However, traditional microscope or X-ray diffraction based crack detection techniques are cost-prohibitive for massive productions. To address this issue, this paper proposes a novel crack sensor design with very little design or testing overhead. It is simply formed by doping the area surrounding a suspicious TSV. By measuring its DC resistances during testing, cracks that grow beyond the doped area can be easily detected. In addition, through empirical studies on crack dynamics in various TSV configurations, we provide deployment guidelines to minimize the number of sensors needed. To the best of our knowledge, this is the first work to propose a macroscale crack detection technique.

In this paper, we present a transient modeling of electromigration (EM) in TSV and TSV-to-wire interfaces in the power delivery network (PDN) of 3D ICs. In particular, we model atomic depletion and accumulation, effective resistance degradation, and full chip-scale PDN lifetime degradation due to EM. Our major focuses are on: (1) time-dependent multi-physics EM modeling approach to model TSVs and connecting wires under the influence of coupled physical phenomenon including electric field, temperature, and stress; (2) time-dependent EM-aware power integrity analysis methodology, which is integrated with the TSV modeling approach to predict long-term IR-drop degradation in full-chip 3D power delivery networks. Our studies show that voids and hillocks grow at various TSV-to-wire interfaces and degrade the effective resistance of TSVs significantly. In addition, our full-chip PDN lifetime analysis shows significant increase in maximum IR drop during lifetime due to EM effects.

This paper presents comprehensive design and analysis results of 3D IC-based low-power stereo matching processors. Our design efforts range from architecture design and verification to RTL-to-GDSII design and sign-off analysis based on GlobalFoundries 130-nm PDK. We conduct comprehensive studies on the area, performance, and power benefits of our 3D IC designs over 2D IC designs. Our 2-tier 3D IC designs attain 43% area, 14% wire length, and 13% power saving over 2D IC designs. We also study a pipeline-based partitioning method shown to be effective at minimizing power consumption and the total number of TSVs while balancing the size of each tier.

In this paper, we study a 3D IC micro-controller implemented with sub-threshold supply for ultra-low power applications. Our study is based on GDSII layouts of a sub-threshold 8052 micro-controller that consumes 3.6μW power running at 20 KHz clock frequency and 0.4V logic supply. Our study confirms that sub-threshold circuits indeed offer a few orders of magnitude power vs performance tradeoff. In addition, our 3D sub-threshold design reduces the footprint area by 78% and wirelength by 33% compared with the 2D counterpart. Our studies also show that thermal and IR drop issues are negligible in this sub-threshold 3D implementation due to its extreme low power operation. Lastly, we demonstrate the low power and high memory bandwidth advantages of many-core 3D sub-threshold circuits.

In this paper, we study the impact of through-silicon-via (TSV) and shallow trench isolation (STI) stress on the timing variations of 3-D IC. We also propose the first systematic TSV-STI-stress-aware timing analysis and show how to optimize layouts for better performance. First, we generate a stress contour map with an analytical radial stress model for TSV. We also develop a stress model for STI from finite element analysis results. Then, depending on geometric relation between TSVs, STI, and transistors, the tensile and compressive stresses are converted to hole and electron mobility variations. Mobility-variation-aware cell library and netlist are generated and incorporated into an industrial engine for timing analysis of 3-D IC. We observe that TSV stress and STI stress interact with each other, and rise and fall time react differently to stress and relative locations with respect to both TSVs and STIs. Overall, TSV-STI-stress-induced timing variations can be as much as ±15% at the cell level. Thus, as an application to layout optimization, we exploit the stress-induced mobility enhancement to improve performance of 3-D ICs. We show that stress-aware layout perturbation could reduce cell delay by up to 23.37% and critical path delay by 6.67% in our test case.

A robust physical design of 3-D IC requires investigation on through-silicon via (TSV). The large temperatures and stress gradients can severely affect TSV delay with large variation. The traditional physical model treats TSV as a resistor with linear electrical-thermal dependence, which ignores the fundamental device physics. In this paper, a physics-based electrical-thermal–mechanical delay model is developed for signal TSVs in 3-D IC. With consideration of liner material and also stress, a nonlinear model is established between electrical delay with temperature and stress. Moreover, sensitivity analysis is performed to relate the reduction of temperature and stress gradients with respect to dummy TSVs insertion. Taking the design of 3-D clock tree as a case study, we have formulated a nonlinear optimization problem for clock-skew reduction. By allocating dummy TSVs to reduce the temperature and stress gradients, the clock skew introduced by signal TSVs and drivers can be minimized. A number of 3-D clock-tree benchmarks are utilized in experiments. We have observed that with the use of dummy TSV insertion, clock skew can be reduced by 61.3% on average when the accurate nonlinear electrical-thermal–mechanical delay model is applied.

In this paper, we propose a fast and accurate chip/package thermomechanical stress co-analysis tool for through-silicon-via (TSV)-based 3-D ICs. We use our tool for full-stack mechanical reliability as well as stress-aware timing analyses. First, we analyze the stress induced by chip/package interconnect elements, i.e., TSV, μ-bump, and package bump. Second, we explore and validate the principle of lateral and vertical linear superposition of stress tensors (LVLS), considering all chip/package elements. The proposed LVLS method greatly reduces the complexity of stress calculation compared with the conventional finite element analysis method with high enough accuracy for full-chip/package-scale stress simulations and reliability analysis. In addition, we build hole and electron mobility variation maps based on LVLS. Finally, we study the mechanical reliability issues and provide full-stack timing analysis results in practical 3-D chip/package designs including wide-I/O and block-level 3-D ICs.

The nano-scale 3-D interconnects available in monolithic 3-D integrated circuit (IC) technology enable ultrahigh density device integration at the individual transistor level. In this paper, we investigate the benefits and challenges of monolithic 3-D integration technology for ultrahigh density logic designs. We first build a 3-D standard cell library for transistor-level monolithic 3-D ICs and model their timing and power characteristics. Then, we explore various interconnect options for monolithic 3-D ICs that improve design quality. Next, we build timing-closed, full-chip GDSII layouts and perform sign-off iso-performance power comparisons with 2-D IC designs. Based on layout simulations, we compare important design metrics such as area, wirelength, timing, and power consumption of transistor-level monolithic 3-D designs with traditional 2-D, gate-level monolithic 3-D, and TSV-based 3-D designs.

The technology of through-silicon vias (TSVs) enables fine-grained integration of multiple dies into a single 3-D stack. TSVs occupy significant silicon area due to their sheer size, which has a great effect on the quality of 3-D integrated chips (ICs). Whereas well-managed TSVs alleviate routing congestion and reduce wirelength, excessive or ill-managed TSVs increase the die area and wirelength. In this paper, we investigate the impact of the TSV on the quality of 3-D IC layouts. Two design schemes, namely TSV co-placement (irregular TSV placement) and TSV site (regular TSV placement), and accompanying algorithms to find and optimize locations of gates and TSVs are proposed for the design of 3-D ICs. Two TSV assignment algorithms are also proposed to enable the regular TSV placement. Simulation results show that the wirelength of 3-D ICs is shorter than that of 2-D ICs by up to 25%.

Imbalanced distribution of workloads across a chip multiprocessor (CMP) constitutes wasteful use of resources. Most existing load distribution and balancing techniques employ very limited hardware support and rely predominantly on software for their operation. This paper introduces IsoNet, a hardware-based conflict-free dynamic load distribution and balancing engine. IsoNet is a lightweight job queue manager responsible for administering the list of jobs to be executed, and maintaining load balance among all CMP cores. By exploiting a micro-network of load-balancing modules, the proposed mechanism is shown to effectively reinforce concurrent computation in many-core environments. Detailed evaluation using a full-system simulation framework indicates that IsoNet significantly outperforms existing techniques and scales efficiently to as many as 1024 cores. Furthermore, to assess its feasibility, the IsoNet design is synthesized, placed, and routed in 45-nm VLSI technology. Analysis of the resulting low-level implementation shows that IsoNet's area and power overhead are almost negligible.

Since re-designing and re-optimizing existing logic, memory, and IP blocks in a 3D fashion significantly increases design cost, near-term three-dimensional integrated circuit (3D IC) design will focus on reusing existing 2D blocks. One way to reuse 2D blocks in the 3D IC design is to first perform 3D floorplanning, insert signal through-silicon vias (TSVs) for 3D inter-block connections, and then route the blocks. In this paper, we propose algorithms (finding signal TSV locations, assigning TSVs to whitespace blocks, and manipulating whitespace blocks) for post-floorplanning signal TSV planning in the block-level 3D IC design. Experimental results show that our signal TSV planner outperforms the state-of-the-art TSV-aware 3D floorplanner by 7% to 38% with respect to wirelength. In addition, our multiple TSV insertion algorithm outperforms a single TSV insertion algorithm by 27% to 37%.

The 3D IC integration using through-silicon-vias (TSV) has gained tremendous momentum recently for industry adoption. However, as TSV involves disruptive manufacturing technologies, new modeling and design techniques need to be developed for 3D IC manufacturability and reliability. In particular, TSVs in 3D IC may cause significant thermal mechanical stress, which not only results in systematic mobility/performance variations, but also leads to mechanical reliability concerns such as interfacial cracking. Meanwhile, the huge dimensional gaps between TSV, on-chip wires, and bonding/packaging all lead to new electromigration concerns. Thus full-chip/package modeling and physical design tools need to be developed to achieve more reliable 3D IC integration. In this paper, we will discuss some key design for manufacturability and reliability challenges and possible solutions for TSV-based 3D IC integration, as well as future research directions.

In this paper, we present an obstacle-aware clock tree synthesis method for through-silicon-via (TSV)-based 3D ICs. A unique aspect of this problem lies in the fact that various types of TSVs become obstacles during 3D clock routing including signal, power/ground, and clock TSVs. Some of these TSVs become placement obstacles, i.e., they interfere with clock buffers and clock TSVs; while other TSVs become routing obstacles, i.e., clock wires cannot route through them. Thus, the key is to perform TSV-induced obstacle-aware 3D clock routing under the following goals: (1) clock TSVs and clock buffers are located while avoiding overlap with placement obstacles; (2) clock wires are routed while avoiding routing obstacles; and (3) clock skew and slew constraints are satisfied. Related experiments show that our TSV-obstacle-aware clock tree does not sacrifice wirelength or clock power too much while avoiding various TSV-induced obstacles.

Thermo-mechanical stress due to TSV fabrication processes is a major concern in 3D integration. TSV stress not only degrades the mechanical reliability of 3D ICs but it also affects the electrical properties, such as electron and hole mobility, of the MOS devices surrounding TSVs. Variations in carrier mobility result in a change in the timing profile of the circuit, which has an impact on delay-fault testing. We show quantitatively using the SDQL metric that test quality is significantly reduced if the test patterns are generated with TSV stress-oblivious circuit models. We evaluate the impact on TSV stress on delay testing by considering layouts for several 3D logic-on-logic benchmarks. The test escape rate is higher for processes with lower yields. Our results also indicate that we can improve the test quality by using TSV-stress aware cell libraries in a conventional ATPG flow with commercial tools, with negligible impact on pattern count. We therefore conclude that any detrimental impact of TSV stress on pattern effectiveness and test quality can be overcome by using stress-aware models for test generation.

In this paper, we propose two methods used in 3D IC placement that effectively exploit the die-to-die thermal coupling in the stack. First, TSVs are spread on each die to reduce the local power density and vertically aligned across dies simultaneously to increase thermal conductivity to the heatsink. Second, we move high-power logic cells to the location that has higher conductivity to the heatsink while moving TSVs in the upper dies so that high-power cells are vertically overlapping below the TSVs. These methods are employed in a force-directed 3D placement successfully and outperform several state-of-the-art placers published in recent literature.

In this work, we propose a fast and accurate chip/package thermomechanical stress and reliability co-analysis tool for TSV-based 3D ICs. We also present a design optimization methodology to alleviate mechanical reliability issues in 3D IC. First, we analyze the stress induced by chip/package interconnect elements, i.e., TSV, μ-bump, and package bump. Second, we explore and validate the principle of lateral and vertical linear superposition of stress tensors (LVLS), considering all chip/package elements. This linear superposition principle is utilized to perform full-chip/package-scale stress simulations and reliability analysis. Finally, we study the mechanical reliability issues in practical 3D chip/package designs including wide-I/O and block-level 3D ICs.

Due to the large geometry of through-silicon-vias (TSVs) and their connections to the power grid, significant current crowding can occur in 3D ICs. Prior works model TSVs and power wire segments as single resistors, which cannot capture the detailed current distribution and may miss trouble spots associated with current crowding. This paper studies DC current crowding and its impact on 3D power integrity. First, we explore the current density distribution within a TSV and its power wire connections. Second, we build and validate effective TSV models for current density distributions. Finally, these models are integrated with global power wires for detailed chip-scale power grid analysis.

Recent innovations in monolithic 3D technology enable much higher-density vertical connections than today's through-silicon-via (TSV)-based technology. In this paper, we investigate the benefits and challenges of monolithic 3D integration technology for ultra high-density logic designs. Based on our layout experiments, we compare important design metrics such as area, wirelength, timing, and power consumption of monolithic 3D designs with the traditional 2D designs. We also explore various interconnect options for monolithic 3D ICs that improve design density and quality. Depending on the interconnect settings of monolithic 3D ICs and the benchmark circuit characteristics, we observe that our two-tier monolithic 3D design provides up to 40% reduced footprint, 27.7% shorter wirelength, 39.7% faster operation, and 9.7% lower power consumption over the 2D counterpart.

Electromigration (EM) has become a key reliability concern for nanometer IC designs. For 3D ICs, higher current density/temperature and TSV-induced thermal mechanical stress further exacerbate the EM issue compared to 2D ICs. In this paper, we analyze the root causes of EM for 3D IC signal nets, with consideration of current density, temperature, and TSV-induced thermal mechanical stress. We develop compact EM models for both DC and AC signal nets using detailed finite-element-analysis (FEA) and build EM library for meantime-to-failure (MTTF). For AC signal nets, we convert AC current into equivalent DC current and model EM with it. One unique property of EM in 3D ICs is that, depending on the current direction, TSV-induced stress may degrade or improve the MTTF, thus routing plays an important role for EM mitigation. We suggest EM-aware routing algorithms for 3D ICs for the first time to our best knowledge, guided by our stress-aware EM modeling. Experimental result shows that our proposed approach improves EM-robustness of 3D IC benchmarks significantly, e.g., 66.4% less EM-violated grids with little sacrifice of conventional routing objectives.

This paper focuses on low-power clock network design for 3D ICs, where through-silicon vias (TSVs) form a regular 2-dimensional array. This TSV array style is shown to be more manufacturable and practical than layouts with TSVs located at irregular spots. However, due to limited TSV resources in TSV arrays, TSV utilization in a 3D clock network significantly affects the final clock power. A straightforward extension on existing works for TSV arrays cannot guarantee power efficiency. Therefore, we develop a decision-tree-based clock synthesis (DTCS) method to generate low-power and reliable clock networks by efficiently exploring the entire solution space for the best TSV array utilization. Our DTCS method has been applied for both gate-level chip-scale 3D clock designs and block-level global clock designs. Experimental results show that our algorithm effectively finds close-to-optimal solutions for power efficiency with skew minimization in short runtime. Our DTCS method achieves up to 13.5% average power reduction with more than 50% fewer TSVs compared with the straightforward extension on the existing algorithm.

Pre-bond testing of TSVs and die logic is a significant challenge and a potential roadblock for 3D integration. BIST solutions introduce considerable die area overhead. Oversized probe pads on TSVs to provide pre-bond test access limit both test bandwidth and TSV density. This paper presents a solution to these problems, allowing a probe card to contact TSVs without the need for probe pads, enabling both TSV and pre-bond scan test. Two possible pre-bond scan test configurations are shown - they provide varying degrees of test parallelism. HSPICE simulations are performed on a logic-on-logic 3D benchmark. Results show that the ratio of the number of probe needles available for test access to the number of pre-bond scan chains determines which pre-bond scan configuration results in the shortest test time. Maximum pre-bond scan-in and scan-out shift-clock speeds are determined for dies in a benchmark 3D design. These clock speeds show that pre-bond scan test can be performed quickly, at a speed that is comparable to scan testing of packaged dies. The maximum clock speed can also be tuned by changing the drive strength of the probe and on-die drivers of the TSV network. Estimates are also provided for peak and average power consumption during pre-bond scan test. On-die area overhead for the proposed method is estimated to be between 1.0% and 2.2% for three dies in the 3D stack.

In order to ensure the correctness of 3D ICs, they need to be tested both before and after their individual dies are bonded. All previous works in the area of 3D IC testing consider only stuck-at fault testing. However, 3D ICs also need to be tested for delay defects. In this work, we present a transition delay test infrastructure that can be used to test a 3D IC both before and after bonding. Furthermore, we present a methodology to test the through silicon vias (TSVs) after bonding, without necessitating regeneration of test patterns. Results show that the overhead involved is negligible. In addition, at-speed testing of circuits can suffer from large IR drop problems. In this paper, we also study the IR drop of 3D ICs during transition delay fault testing. We study how different configurations of probe pads affect the pre-bond IR drop. We also study how this IR drop changes from the pre-bond to the post-bond case.

In this paper, we propose an efficient and accurate full-chip thermomechanical stress and reliability analysis tool and design optimization methodology to alleviate mechanical reliability issues in 3-D integrated circuits (ICs). First, we analyze detailed thermomechanical stress induced by through-silicon vias in conjunction with various associated structures such as landing pad and dielectric liner. Then, we explore and validate the linear superposition principle of stress tensors and demonstrate the accuracy of this method against detailed finite element analysis simulations. Next, we apply this linear superposition method to full-chip stress simulation and a reliability metric named the von Mises yield criterion. Finally, we propose a design optimization methodology to mitigate the mechanical reliability problems in 3-D ICs. Our numerical experimental results demonstrate the effectiveness of the proposed methodology.

This paper presents a design methodology for robust and low-energy clock networks for ultralow voltage (ULV) circuits. We show that both clock slew and skew play important roles in achieving high maximum operating frequency$(F_{\max})$and low clock energy in ULV circuits. In addition, clock networks in ULV circuits are highly sensitive to process variations. We propose a variation-aware methodology that controls both clock skew and slew to maximize$F_{\max}$and minimize clock power. In addition, we implement dynamic programming (DP)-based ULV clock routing and buffering methods (deferred merging and embedding) for deterministic and statistical conditions. Experimental results show that our clock network design method achieves lower energy (more than 20% savings) at comparable or even higher$F_{\max}$compared with the existing methods.

3-D integration has the potential to increase performance and decrease energy consumption. However, there are many unsolved issues in the design of these systems. In this work we study the design of 3-D power supply networks and demonstrate a technique specific to 3-D systems that improves IR-drop and dynamic noise over a straightforward extension of traditional design techniques. Previous work in 3-D power delivery network design has simply extended 2-D techniques by treating through-silicon vias (TSVs) as extensions of the C4 bumps. By exploiting the smaller size and much higher interconnect density possible with TSVs we demonstrate significant reduction of nearly 50% in the IR-drop and 42% in the dynamic noise of our large-scale 3-D design. Simulations also show that a 3-tier stack with the distributed TSV topology actually lowers IR-drop by 21% and dynamic noise by 32% over a non-3-D system with less power dissipation. We analyze the power distribution network of an envisioned 1000-core processor with 30 stacked dies and show scaling trends related to both increased stacking and power distribution TSVs. Finally, we examine several techniques for minimizing IR-drop and dynamic noise and their effects on our large-scale 3-D system.

3D integration has new manufacturing and design challenges such as timing corner mismatch between tiers and device variation due to Through Silicon Via (TSV) induced stress. Timing corner mismatch between tiers is caused because each tier is manufactured in independent process. Therefore, inter-die variation should be considered to analyze and optimize for paths spreading over several tiers. TSV induced stress is another challenge in 3D Clock Tree Synthesis (CTS). Mobility variation of a clock buffer due to stress from TSV can cause unexpected skew which degrades overall chip performance. In this paper, we propose clock tree design methodology with the following objectives: (a) to minimize clock period variation by assigning optimal z-location of clock buffers with an Integer Linear Program (ILP) formulation, (b) to prevent unwanted skew induced by the stress. In the results, we show that our clock buffer tier assignment reduces clock period variation up to 34.2%, and the most of stress-induced skew can be removed by our stress-aware CTS. Overall, we show that performance gain can be up to 5.7% with our robust 3D CTS.

In this work, we propose an efficient and accurate full-chip thermo-mechanical stress and reliability analysis tool and design optimization methodology to alleviate mechanical reliability issues in 3D ICs. First, we analyze detailed thermo-mechanical stress induced by TSVs in conjunction with various associated structures such as landing pad and dielectric liner. Then, we explore and validate the use of the linear superposition principle of stress tensors and demonstrate the accuracy of this method against detailed finite element analysis (FEA) simulations. Next, we apply this linear superposition method to full-chip stress simulation and a reliability metric named the von Mises yield criterion. Finally, we propose a design optimization methodology to mitigate the mechanical reliability problems in 3D ICs. Our experimental results demonstrate the effectiveness of our methodology.

This paper studies TSV-to-TSV coupling in 3D ICs. A full-chip SI analysis flow is proposed based on the proposed coupling model. Analysis results show that TSVs cause significant coupling noise and timing problems despite that TSV count is much smaller compared with the gate count. Two approaches are proposed to alleviate TSV-to-TSV coupling, namely TSV shielding and buffer insertion. Analysis results show that both approaches are effective in reducing the TSV-caused-coupling and improving timing.

3D integration has the potential to increase performance and decrease energy consumption. However, there are many unsolved issues in the design of these systems. In this work we study the design of many-tier (more than 4 tiers stacked) 3D power-supply networks and demonstrate a technique specific to 3D systems that improves IR-drop over a straightforward extension of traditional design techniques. Previous work in 3D power delivery network design has simply extended 2D techniques by treating through-silicon vias (TSVs) as extensions of the C4 bumps. By exploiting the smaller size and much higher interconnect density possible with TSVs we demonstrate significant reduction of nearly 50% in the IR-drop of our 3D design. Simulations also show that a 3-tier stack with the distributed TSV topology actually lowers IR-drop by 20% over a non-3D system with less power dissipation. Finally, we analyze the power distribution network of an envisioned 1000-core processor with 30 stacked dies and show scaling trends related to both increased stacking and power distribution TSVs. Our 3D analysis technique is validated using commercial-grade sign-off IR-drop software from a major EDA vendor.

In this work, we propose an efficient and accurate full-chip through-silicon-via (TSV) interfacial crack analysis flow and design optimization methodology to alleviate TSV interfacial crack problems in 3D ICs. First, we analyze TSV interfacial crack at TSV/dielectric liner interface caused by TSV-induced thermo-mechanical stress. Then, we explore the impact of TSV placement in conjunction with various associated structures such as landing pad and dielectric liner on TSV interfacial crack. Next, we present a full-chip TSV interfacial crack analysis methodology based on design of experiments (DOE) and response surface method (RSM). Finally, we propose a design optimization methodology to mitigate the mechanical reliability problems in 3D ICs.

Electromigration (EM) is a critical problem for interconnect reliability of modern integrated circuits (ICs), especially as the feature size becomes smaller. In three-dimensional (3D) IC technology, the EM problem becomes more severe due to drastic dimension mismatches between metal wires, through silicon vias (TSVs), and landing pads. Meanwhile, the thermo-mechanical stress due to the TSV can also cause reduction in the failure time of wires. However, there is very little study on EM issues that consider TSVs in 3D ICs. In this paper, we show the impact of TSV stress on EM failure time of metal wires in 3D ICs. We model the impact of TSV on stress variation in wires. We then perform detailed modeling of the impact of stress on EM failure time of metal wires. Based on our analysis, we build a detailed library to predict the failure time of a given wire based on current density, temperature and stress. We then propose a method to perform fast full-chip simulation, to determine the various EM related hot-spots in the design. We also propose a simple routing-blockage scheme to reduce the EM related failures near the TSVs, and see its impact on various metrics.

3D integration is a promising new technology for tightly integrating multiple active silicon layers into a single chip stack. Both the integration of heterogeneous tiers and the partitioning of functional units across tiers leads to significant improvements in functionality, area, performance, and power consumption. Managing the complexity of 3D design is a significant challenge that will require a system-on-chip approach, but the application of SOC design to 3D necessitates extensions to current test methodology. In this paper, we propose extending test wrappers, a popular SOC DFT technique, into the third dimension. We develop an algorithm employing the Best Fit Decreasing and Kernighan-Lin Partitioning heuristics to produce 3D wrappers that minimize test time, maximize reuse of routing resources across test modes, and allow for different TAM bus widths in different test modes. On average the two variants of our algorithm reuse 93% and 92% of the test wrapper wires while delivering test times of just 0.06% and 0.32% above the minimum.

Pre-bond testing of 3D ICs improves yield by preventing bad dies and/or wafers from being used in the final 3D stack. However, pre-bond testing is challenging because it requires special scan chains and power delivery mechanism. Any 3D scan chains that traverse multiple dies will be fragmentized in each individual die during pre-bond testing. In this paper we study the scan chain and power delivery network synthesis for pre-bond testing of 3D ICs. The testing of individual dies is facilitated by the addition of dedicated probe pads for power delivery and scan IO as a form of design-for-testing. We investigate the impact of scan-chain Through-Silicon-Vias (TSVs) on power consumption and voltage drop. We also study the requirements of power probe pads for power delivery during pre-bond structural test.

Heat removal and power delivery have become two major reliability concerns in 3-D integrated circuit (IC) technology. To alleviate thermal problem, two possible solutions have been proposed: thermal-through-silicon-vias (T-TSVs) and micro-fluidic channel (MFC)-based cooling. In case of power delivery, a complex power distribution network is required to deliver currents reliably to all parts of the 3-D IC while suppressing the power supply noise to an acceptable level. However, these thermal and power networks pose major challenges in signal routability and congestion. This is because signal, power, and thermal interconnects are all competing for routing space, and the related TSVs interfere with gates and wires in each die. We present a co-optimization methodology for signal, power, and thermal interconnects in 3-D ICs based on design of experiments (DOE) and response surface methodology (RSM). The goal of our holistic approach is to improve signal, thermal, and power noise metrics and to provide fast and accurate design space exploration for early design stage. We also provide an in-depth comparison between T-TSV versus MFC-based cooling method and discuss how to employ DOE and RSM techniques to co-optimize the interconnects. Our DOE-based optimization found the optimal design point with less effort than a gradient search-based optimization.

In this paper, we analyze the effect of clock slew in subthreshold circuits. Specifically, we address the issue that variations in clock slew at the register control can cause serious timing violations. We show that clock slew variations can cause frequency targets to deviate by as much as 28% from the design goals. Based on these observations, we recognize the importance of clock slew control in subthreshold circuits. We propose a systematic approach to design the clock tree for subthreshold circuits to reduce the clock slew variations while minimizing the energy dissipation in the tree. The combined approach, including the wire sizing and dynamic nodal capacitance control, can achieve better slew control (and better timing control) at lower energy in subthreshold circuits.

Pre-bond testing of 3-D stacked integrated circuits (ICs) involves testing each individual die before bonding. The overall yield of 3-D ICs improves with pre-bond testability because manufacturers can avoid stacking defective dies with good ones. However, pre-bond testability presents unique challenges to 3-D clock tree design. First, each die needs a complete 2-D clock tree to enable pre-bond test. Second, the entire 3-D stack needs a complete 3-D clock tree for post-bond test and operation. In the case of a two-die stack, a straightforward solution is to have two complete 2-D clock trees connected with a single through-silicon-via (TSV). We show that this solution suffers from long wirelength (WL) and high clock power consumption. Our algorithm improves on this solution, minimizes the overall WL and clock power consumption, and provides both pre-bond testability and post-bond operability with minimum skew and constrained slew. Compared with the single-TSV solution, SPICE simulation results show that our multi-TSV approach significantly reduces the clock power by up to 15.9% for two-die and 29.7% for four-die stacks. In addition, the WL is reduced by up to 24.4% and 42.0%.

In this article, we propose a design methodology using two complementary techniques to address high-frequency inductive noise in the early design phase of a microprocessor. First, we propose a noise-aware floorplanning technique that uses microarchitectural profile information to create noise-aware floorplans. Second, we present the design of a dynamic inductive-noise controlling mechanism at the microarchitectural level, which limits the on-die current demand within predefined bounds, regardless of the native power and current characteristics of running applications. By dynamically monitoring the access patterns of microarchitectural modules, our mechanism can effectively limit simultaneous switching activity of close-by modules, thereby leveling voltage ringing at local power-pins. Compared to prior art, our di/dt alleviation technique is the first that takes the processor's floorplan, as well as its power-pin distribution, into account to provide a finer-grained control with minimal performance degradation. Based on the evaluation results using 2D floorplans, we show that our techniques can significantly improve inductive noise induced by current demand variation and reduce the average current variability by up to 7 times, with an average performance overhead of 4.0%. In addition, our floorplan reduces the noise margin violations using our noise-aware floorplan by an average of 56.3% while reducing the decap budget by 28%.

In this paper, we study the clock tree sizing problem for thermal-aware skew minimization under power and thermal budgets. Clock wire/buffer sizing affects not only the delay/skew, but also the power dissipation of the clock tree. This effect in turn triggers changes in thermal distribution, making re-computation of the delay/skew necessary. Thus, the interaction among skew, power, and temperature is highly complicated if tied with clock wire/buffer sizing. In order to efficiently combat the time-varying nature of underlying thermal profile, we focus on two kinds of skew, depending on the number of thermal profiles given: skew value and skew range. The former refers to the skew value computed under a single steady-state thermal profile, whereas the latter refers to the skew range computed based on multiple thermal profiles. Our thermal-aware sequential-linear-programming approach maintains near-zero skew value and narrow skew range while keeping the power dissipation and temperature under the given budgets.

In this paper, three effective design techniques are presented to effectively reduce the clock power consumption and slew of the 3D clock distribution network: (1) controlling the bound of through-silicon-vias (TSVs) used in between adjacent dies, (2) controlling the maximum load capacitance of the clock buffer, (3) adjusting the clock source location in the 3D stack. We discuss how these design factors affect the overall wirelength, clock power, slew, skew, and routing congestion in the practical 3D clock network design. SPICE simulation indicates that: (1) a 3D clock tree with multiple TSVs achieves up to 31% power saving, 52% wirelength saving and better slew control as compared with the single-TSV case; (2) by placing the clock source on the middle die in the 3D stack, an additional 7.7% power savings, 9.2% wirelength savings, and 33% TSV savings are obtained compared with the clock source on the topmost die. This work aims at helping designers construct reliable low-power and low-slew 3D clock network by making the right decisions on TSV usage, clock buffer insertion, and clock source placement.

As the geometry shrinking faces severe limitations, 3D wafer stacking with through silicon via (TSV) has gained interest for future SOC integration. Since TSV fill material and silicon have different coefficients of thermal expansion (CTE), TSV causes silicon deformation due to different temperatures at chip manufacturing and operating. The widely used TSV fill material is copper which causes tensile stress on silicon near TSV. In this paper, we propose systematic TSV stress aware timing analysis and show how to optimize layout for better performance. First, we generate a stress contour map with an analytical radial stress model. Then, the tensile stress is converted to hole and electron mobility variations depending on geometric relation between TSVs and transistors. Mobility variation aware cell library and netlist are generated and incorporated in an industrial timing engine for 3D-IC timing analysis. It is interesting to observe that rise and fall time react differently to stress and relative locations with respect to TSVs. Overall, TSV stress induced timing variations can be as much as ± 10% for an individual cell. Thus as an application for layout optimization, we can exploit the stress-induced mobility enhancement to improve timing on critical cells. We show that stress-aware perturbation could reduce cell delay by up to 14.0% and critical path delay by 6.5% in our test case.

Through-silicon via (TSV) fabrication causes tensile stress around TSVs which results in significant carrier mobility variation in the devices in their neighborhood. Keep-out zone (KOZ) is a conservative way to prevent any devices/cells from being impacted by the TSV-induced stress. However, owing to already large TSV size, large KOZ can significantly reduce the placement area available for cells, thus requiring larger dies which negate improvement in wirelength and timing due to 3D integration. In this paper, we study the impact of KOZ dimension on stress, carrier mobility variation, area, wirelength, and performance of 3D ICs. We demonstrate that, instead of requiring large KOZ, 3D-IC placers must exploit TSV stress-induced carrier mobility variation to improve the timing and area objectives during placement. We propose a new TSV stress-driven force-directed 3D placement that consistently provides placement result with, on average, 21.6% better worst negative slack (WNS) and 28.0% better total negative slack (TNS) than wirelength-driven placement.

In this paper we present a test structure and design methodology for testing, characterization, and self-repair of TSVs in 3D ICs. The proposed structure can detect the signal degradation through TSVs due to resistive shorts and variations in TSV. For TSVs with moderate signal degradations, the proposed structure reconfigures itself as signal recovery circuit to improve signal fidelity. The paper presents the design of the test/recovery structure, the test methodologies, and demonstrates its effectiveness through stand alone simulations as well as in a full-chip physical design of a 3D IC.

In 3D integrated circuits through silicon vias (TSVs) are used to connect different dies stacked on top of each other. These TSV occupy silicon area and have significantly larger area than regular gates. In this paper, we address two critical aspects of TSV management in 3D designs. First, we address the problem of how many TSVs to add in a design. Since TSVs occupy significant silicon area, a general tendency has been to use a minimum number of TSVs in 3D circuits. We show that such an approach does not give us the best possible result. Second, we address the problem of TSV insertion. Because TSVs occupy silicon area, their location is decided during the placement stage of 3D design. However, we show that this is not the best possible stage for TSV insertion. We propose a change in the physical design flow for 3D integrated circuits to address the limitations of existing TSV placement methodology. All our algorithms are integrated with commercial tools, and our results are validated based on actual GDSII layouts. Our experimental results show the effectiveness of our methods.

Multi-core microarchitectures require a careful balance between many competing objectives to achieve the highest possible performance. Integrated early analysis is the consideration of all of these factors at an early stage. Toward this goal, this work presents the first adaptive multi-granularity multi-core microarchitecture-level floorplanner that simultaneously optimizes temperature and performance, and considers memory bus length. We include simultaneous optimization at both the module-level and the core/cache-bank level. Related experiments show that our methodology is effective for optimizing multi-core architectures.

Heat removal and power delivery are two major reliability concerns in the 3D stacked IC technology. Liquid cooling based on micro-fluidic channels is proposed as a viable solution to dramatically reduce the operating temperature of 3D ICs. In addition, designers use a highly complex hierarchical power distribution network in conjunction with decoupling capacitors to deliver currents to all parts of the 3D IC while suppressing the power supply noise to an acceptable level. These so called silicon ancillary technologies, however, pose major challenges to routing completion and congestion. These thermal and power/ground interconnects together with those used for signal delivery compete with one another for routing resources including various types of Through-Silicon-Vias (TSVs). This paper presents the work on routing with these interconnects in 3D: signal, power, and thermal networks. We demonstrate how to consider various physical, electrical, and thermo-mechnical requirements of these interconnects to successfully complete routing while addressing various reliability concerns.

Decoupling capacitors (decaps) are typically used to reduce the noise in the power supply network. Because the delay of gates and interconnects is affected by the supply voltage level, decaps can be used to improve the circuit performance as well. In this paper, we present the analytical delay model under IR drop, Ldi/dt noise, and decaps to study how decaps affect both the gate and interconnect delay. Given a floorplanning solution, we study how to allocate the whitespace for decap insertion so that the delay is minimized under the given noise and area constraint. We employ the Sequential Linear Programming method to solve the non-linear whitespace allocation problem. Our experimental results show that intelligent decap allocating decap makes further delay reduction possible without adding any additional decap.

Through-Silicon-Via (TSV) is the enabling technology for the fine-grained 3D integration of multiple dies into a single stack. These TSVs occupy non-negligible silicon area because of their sheer size. This significant silicon area occupied by the TSVs and the interconnections made to the TSVs greatly affect area, power, performance, and reliability of 3D IC layouts. Well-managed TSVs alleviate congestion, reduce wirelength, and improve performance, whereas excessive TSVs not only increase the die area, but also have negative impact on many design objectives. In this paper, we study the impact of TSV on various aspects of 3D layouts. We use GDSII layouts of 2D and 3D designs, and thoroughly compare the pros and cons of TSV usage. We propose a new force-directed 3D gate-level placement that efficiently handles TSVs. In addition, we present an algorithm that assigns TSVs to nets to complete routing that involves TSVs. This algorithm, together with our 3D placer, is integrated into a commercial P&R tool to generate fully validated GDSII layouts. Our experiments based on synthesized benchmarks indicate that our algorithms help generate GDSII layouts of 3D designs that are optimized in terms of area, wirelength, and metal layer count.

Heat removal and power delivery have become two major reliability concerns in 3D stacked IC technology. For thermal problem, two possible solutions exist: thermal-through-silicon-vias (T-TSVs) and micro-fluidic channel (MFC) based liquid cooling. In case of power delivery, a highly complex power distribution network is required to deliver currents reliably to all parts of the 3D stacked IC while suppressing the power supply noise to an acceptable level. However, these thermal and power networks pose major challenges in signal routability and congestion. This is because the signal, power, and thermal interconnects are all competing for routing space. In this paper, we present a co-optimization methodology for the signal, power, and thermal interconnects for 3D stacked ICs based on design of experiments (DOE) and response surface method (RSM). The goal is to improve performance, thermal, noise, and congestion metrics with our holistic approach. We also provide in-depth comparison between T-TSV and MFC based cooling method and discuss how to employ DOE and RSM to best co-optimize the multi-functional interconnects simultaneously.

Pre-bond testing of 3D stacked ICs involves testing individual dies before bonding. The overall yield of 3D ICs improves with pre-bond testability because designers can avoid stacking defective dies with good ones. However, pre-bond testability presents unique challenges to 3D clock tree design. First, each die needs a com-plete 2D clock tree for the pre-bond testing. In addition, the en-tire 3D stack needs a complete 3D clock tree for post-bond testing and normal operations. In the case of two-die stack, a straightfor-ward solution is to have two complete 2D clock trees connected with a single Through-Silicon-Via (TSV). We show that this solu-tion suffers from long wirelength and high clock power consump-tion. Instead, our algorithm minimizes the overall wirelength and clock power consumption while providing the pre-bond testability and post-bond operability under given skew and slew constraints. Compared with the single-TSV solution, SPICE simulation results show that our multi-TSV approach significantly reduces the clock power by up to 15.9 % for two-die and 29.7 % for four-die stack. In addition, the wirelength reduction is up to 24.4 % and 42.0%. 1.

In the paper, we analyze the effect of clock slew in subthreshold circuits. Specifically, we address the issue that variations in clock slew at the register control can cause serious timing violations. We show that clock slew variations can cause latch timing metrics such as setup, hold and clock-to-q times to deviate by 90% from the design goals. Based on these observations, we recognize the importance of clock slew control in subthreshold circuits. We propose a systematic approach to design the clock tree for subthreshold circuits to reduce the clock slew variations while minimizing the power dissipation in the tree. We show that a tighter nodal capacitance control is necessary to control the slew in a subthreshold clock tree, which can increase the power dissipation. Recognizing that the wire resistances have a negligible effect in subthreshold circuits, we show proper wire sizing is necessary to reduce the clock power. Finally, we propose a dynamic nodal capacitance control technique that allows larger slew at the earlier nets of the tree while controlling it more aggressively near the sink nodes. The combined approach, including the wire sizing and dynamic nodal capacitance control, can achieve better slew control (and better timing control) at lower power in subthreshold circuits.

In this paper, we present a performance and thermal-aware Steiner routing algorithm for three-dimensional (3-D) stacked integrated circuits. Our algorithm consists of two steps: tree construction and tree refinement. Our tree construction algorithm builds a delay-oriented Steiner tree under a given thermal profile. We show that our 3-D tree construction involves minimization of two-variable Elmore delay function. In our tree refinement algorithm, we reposition the through-silicon-vias (TSVs) used in existing Steiner trees while preserving the original routing topology for further thermal optimization under a performance constraint. We employ a novel scheme to relax the initial nonlinear programming formulation to integer linear programming and consider all TSVs from all nets simultaneously. Our tree construction algorithm outperforms the popular 3-D maze routing by 52% in terms of performance at the cost of 15% wirelength and 6% TSV count increase for four-die stacking. In addition, our TSV relocation results in 9% maximum-temperature reduction at no additional area cost. We also provide extensive experimental results, including the following: (1) the wirelength and delay distribution of various types of 3-D interconnects; (2) the impact of TSVRCparasitics on routing and TSV relocation; and (3) the impact of various bonding styles on routing and TSV relocation. Last, we provide results on two-die stacking.

In this paper, we present a novel design methodology to combat the ever-aggravating high frequency power supply noise (di/dt) in modern microprocessors. Our methodology integrates microarchitectural profiling for noise-aware floorplanning, dynamic runtime noise control to prevent unsustainable noise emergencies, as well as decap allocation; all to produce a design for the average-case current consumption scenario. The dynamic controller contributes a microarchitectural technique to eliminate occurences of the worst-case noise scenario thus our method focuses on average-case noise behavior.

In this paper we present the first bus-aware microarchitectural floorplanning. Our goal is to study the impact of bus routability on other important floorplanning objectives including area, performance, power, and thermal. We developed a fast performance-aware bus routing algorithm, which is integrated into the floorplanning engine to ensure routability while optimizing other conflicting objectives. Our related experiments performed on high performance processors show that we obtain 100% routability at the cost of minimal increase on area, performance, and power objectives under thermal constraint.

In this paper, we study the buffered clock tree synthesis problem under thermal variations for 3D IC technology. Our major contribution is the Balanced Skew Theorem, which provides a theoretical background to efficiently construct a buffered 3D clock tree that minimizes and balances the skew values under two distinct non-uniform thermal profiles. Our clock tree synthesis algorithm named BURITO (buffered clock tree with thermal optimization) first constructs a 3D abstract tree under the wirelength vs via-congestion tradeoff. This abstract tree is then embedded, buffered, and refined under the given non-uniform thermal profiles so that the temperature-dependent skews are minimized and balanced. Experimental results show that our algorithms significantly reduce and perfectly balance clock skew values with minimal wirelength overhead.

Circuit and processor designs will continue to increase in complexity for the foreseeable future. With these increasing sizes comes the use of wide buses to move large amounts of data from one place to another. Bus routing has therefore become increasingly important. In this paper, we present a new bus routing algorithm that globally optimizes both the floorplan and the bus routes themselves. Our algorithm is based on creating a range of feasible bus positions and then using Linear Programming to optimally solve for bus locations. We present this algorithm for use in microarchitectures and explore several different optimization objectives, including performance, floorplan area, and power consumption. Our results demonstrate that this algorithm is effective for efficiently generating feasible routes for complex modern designs and provides better results than previous approaches.

In this paper we present the Statistical Retiming-based Timing Analysis (SRTA) algorithm. The goal is to compute the timing slack distribution for the nodes in the timing graph and identify the statistically critical paths under retiming, which are the paths with a high probability of becoming timing-critical after retiming. SRTA enables the designers to perform circuit optimization on these paths to reduce the probability of them becoming timing bottleneck if the circuit is retimed as a post-process. We provide a comparison among static timing analysis (= STA), statistical timing analysis (= SSTA), retiming-based timing analysis (= RTA), and our statistical retiming-based timing analysis (SRTA). Our results show that the placement optimization based on SRTA achieves the best performance results.

This paper proposes noise-direct, a design methodology for power integrity aware floorplanning, using microarchitectural feedback to guide module placement. Stringent power constraints have led microprocessor designers to incorporate aggressive power saving techniques such as clock-gating, that place a significant burden on the power delivery network. While the application of extensive clock-gating can effectively reduce power consumption, unfortunately, it can also induce large inductive noise (di/dt), resulting in signal integrity and reliability issues. To combat these problems, processors are usually designed for the worst-case current consumption scenario using adequate supply voltage and decoupling capacitances. To tackle high-frequency inductive noise and potential IR drops, we propose a novel design methodology that integrates microarchitectural profiling feedback into the floorplanning process. We present two microarchitectural metrics to quantify the noise susceptibility of a module:self weighting and correlation weighting. By using these metrics in a force-directed floorplanning algorithm to assign power pin affinity to modules, we can quickly converge to a design for average-case current consumption. By designing for the average-case and employing dynamic di/dt control for the worst-case, we can ensure that a chip is noise-tolerant without exceeding decap budget constraints. Our observations showed that certain functional modules in a processor exhibit consistent and highly correlated switching activity, that can be used to guide module placement distance from power pins. The experimental results demonstrate that the force-directed floorplanning technique can effectively suppress supply noise experienced by modules, reduce the total number of supply-noise margin violations, and achieve a floor-plan with considerably lower IR drop, as compared to a wire-length driven floorplan.

Modern advances in reconfigurable analog technologies are allowing field-programmable analog arrays (FPAAs) to dramatically grow in size, flexibility, and usefulness. This paper presents rapid prototyping results of a bandpass filter as a sample analog circuit using our floating-gate based large-scale FPAA. A major source of parasitics introduced during the circuit mapping process is interconnect switches used for routing. Our goal is to obtain models of the mapped circuits that can be simulated using SPICE in order to observe the impact of interconnect parasitics on the relevant analog metrics. Our results indicate that the mapped analog circuits obtain desired responses even with interconnect parasitics, clearly demonstrating the practicality of our FPAA.

In this paper, we present the first work on the Steiner routing for 3D stacked ICs. In the 3D Steiner routing problem, the pins are located in multiple device layers, which makes it more general than its 2D counterpart. Our algorithm consists of two steps: tree construction and tree refinement. Our tree construction algorithm builds a delay-oriented Steiner tree under a given thermal profile. We show that thermal-aware 3D tree construction involves the minimization of two-variable Elmore delay function. In our tree refinement algorithm, we reposition the through-vias while preserving the original routing topology for further thermal optimization under performance constraint. We employ a novel scheme to relax the initial NLP formulation to ILP and consider all through-vias from all nets simultaneously. Our related experiments show the effectiveness of our proposed solutions.

Physical layout generation of RF embedded passive design is not an easy task since the response of a given layout is tightly coupled with the response of the individual components and the effect of interconnect parasitics. In this paper we propose a methodology for automatic layout generation of embedded passive RF circuits. We make use of circuit models to represent and optimize a given layout and use non-linear optimization at various stages of the methodology to obtain the desired goals. Full-wave EM simulations is completely out of the design loop, so our methodology significantly reduces the design time for RF embedded passive circuits. The proposed approach has been used successfully to generate layout for band-pass filters of varying sizes.

One of the biggest challenges in 3D stacked IC design is heat dissipation. Incorporating thermal vias is a promising method for reducing the temperatures of 3D ICs. The bonding styles between device layers impose certain restrictions to where thermal vias may be inserted. This paper presents a whitespace redistribution algorithm that takes bonding style into consideration to improve thermal via placement, which in turn reduces temperature.

This paper presents the first multiobjective microarchitectural floorplanning algorithm for high-performance processors implemented in two-dimensional (2-D) and three-dimensional (3-D) ICs. The floorplanner takes a microarchitectural netlist and determines the dimension as well as the placement of the functional modules into single- or multiple-device layers while simultaneously achieving high performance and thermal reliability. The traditional design objectives such as area and wirelength are also considered. The 3-D floorplanning algorithm considers the following 3-D-specific issues: vertical overlap optimization and bonding-aware layer partitioning. The hybrid floorplanning approach combines linear programming and simulated annealing, which is shown to be very effective in obtaining high-quality solutions in a short runtime under multiobjective goals. This paper provides comprehensive experimental results on making tradeoffs among performance, thermal, area, and wirelength for both 2-D and 3-D ICs

Decoupling capacitors (decaps) are a popular means for reducing power-supply noise in integrated circuits. Since the decaps are usually inserted in the whitespace of the device layer, decap management during the floorplanning stage is desirable. However, a well-known existing work only allows the blocks to utilize the adjacent whitespace. In order to overcome this limit, we devise the effective-decap-distance model to analyze how functional blocks are affected by nonneighboring decaps. In addition, we propose a generalized network-flow-based algorithm to allocate the whitespace to the blocks and determine the oxide thicknesses for the decaps to be implemented in the whitespace. Experimental results show that our decap allocation and sizing methods can significantly reduce decap budget and leakage power with a small increase in area and wire length when integrated into 2D and 3D floorplanners.

introduction Introduction to special issue on demonstrable software systems and hardware platforms Authors: Sung Kyu Lim Atlanta, Georgia Atlanta, GeorgiaView Profile , Massoud Pedram Los Angeles, California Los Angeles, CaliforniaView Profile Authors Info & Claims ACM Transactions on Design Automation of Electronic SystemsVolume 12Issue 3August 2007 Article No.: 20pp 1–3https://doi.org/10.1145/1255456.1255457Published:22 May 2008Publication History 1citation283DownloadsMetricsTotal Citations1Total Downloads283Last 12 Months1Last 6 weeks1 Get Citation AlertsNew Citation Alert added!This alert has been successfully added and will be sent to:You will be notified whenever a record that you have chosen has been cited.To manage your alert preferences, click on the button below.Manage my AlertsNew Citation Alert!Please log in to your account Save to BinderSave to BinderCreate a New BinderNameCancelCreateExport CitationPublisher SiteGet Access

Process variations in digital circuits make sequential circuit timing validation an extremely challenging task. In this paper, a Statistical Bellman-Ford (SBF) algorithm is proposed to compute the longest path length distribution for directed graphs with cycles. Our SBF algorithm efficiently computes the statistical longest path length distribution if there exist no positive cycles or detects one if the circuit is likely to have a positive cycle. An important application of SBF is Statistical Retiming based Timing Analysis (SRTA), where SBF is used to check for the feasibility of a given target clock period distribution for retiming. Our gate and wire delay distribution model considers several high-impact intra-die process parameters and accurately captures the spatial and reconvergent path correlations. The Monte Carlo simulation is used to validate the accuracy of our SBF algorithm. To the best of our knowledge, this is the first paper that propose the statistic version of the longest path algorithm for sequential circuits

In this paper, we present the first multi-objective microarchitectural floorplanning algorithm for designing highperformance, high-reliability processors in the early design phase. Our floorplanner takes a microarchitectural netlist and determines the placement of the functional modules while simultaneously optimizing for performance and thermal reliability. The traditional design objectives such as area and wirelength are also considered. Our multi-objective hybrid floorplanning approach combining Linear Programming and Simulated Annealing is shown to be fast and effective in obtaining high-quality solutions. We evaluate the trade-off of performance, temperature, area, and wirelength and provide comprehensive experimental results.

Optical interconnects enable faster signal propagation with virtually no crosstalk. In addition, wavelength division multiplexing allows a single waveguide to be shared among multiple interconnects. This paper proposes efficient algorithms for the construction of timing and congestion-driven waveguides considering the optical resource constraints. We develop the first optical router for System-on-Packages (SOPs), which reduce electrical wirelength by 11% and improve performance by 23%, when a single optical layer is introduced for every placement layer.

3D circuits have the potential to improve performance over traditional 2D circuits by reducing wirelength and interconnect delay. One major problem with 3D circuits is that their higher device density due to reduced footprint area leads to greater temperatures. Thermal vias are a potential solution to this problem. This paper presents a thermal via insertion algorithm that can be used to plan thermal via locations during floorplanning. The thermal via insertion algorithm relies on a new thermal analyzer based on random walk techniques. Experimental results show that, in many cases, considering thermal vias during floorplanning stages can significantly reduce the temperature of a 3D circuit

Decoupling capacitor (decap) is a popular means to reduce power supply noise in integrated circuits. Since the decaps are usually inserted in the whitespace of the device layer, decap management during the floorplanning stage is desirable. In this paper, we devise the Effective Decap Distance model to analyze how functional blocks are affected by non-neighboring decaps. In addition, we propose a generalized network flow-based algorithm to allocate the whitespace to the blocks and determine the oxide thicknesses for the decaps to be implemented in the whitespace. Experimental results show that our decap allocation and sizing methods can significantly reduce decap budget and leakage power with a small increase in area and wirelength when integrated into 2D and 3D floorplanners.

The integration of retiming and simultaneous supply/threshold voltage scaling has a potential to enable more rigorous total power reduction. However, such integration is a highly complex task due to its enormous solution space. This paper presents the first algorithm that performs retiming and simultaneous supply/threshold voltage scaling. In our three-step approach, low power retiming is first performed to reduce the clock period while taking the FF delay/power into consideration. Next, the subsequent voltage scaling makes the best possible supply/threshold voltage assignment under the given clock period constraint set by the retiming. Finally, a post-process further refines the voltage scaling solution by exploiting the remaining timing slack in the circuit. Related experiments show that our min-FF retiming plus simultaneous Vdd/Vth scaling approach reduces the total power consumption by 34% on average compared to the existing max-FF retiming plus Vdd Scaling approach.

Power delivery is a growing reliability concern in microprocessors as the industry moves toward feature-rich, power-hungrier designs. To battle the ever-aggravating power consumption, modern microprocessor designers or researchers propose and apply aggressive power-saving techniques in the form of clock-gating and/or power-gating in order to operate the processor within a given power envelope. These techniques, however, often lead to high-frequency current variations, which can stress the power delivery system and jeopardize reliability due to inductive noise (L(di/dt)) in the power supply network. To counteract these issues, modern microprocessors are designed to operate under the worst-case current assumption by deploying adequate decoupling capacitance. With the trend of lower supply voltage and increased leakage power and current consumption, designing a processor for the worst case is becoming less appealing. In this paper, we propose a new dynamic inductive-noise controlling mechanism at the microarchitectural level that limit the on-die current demand within predefined bounds, regardless of the native power and current characteristics of running applications. By dynamically monitoring the access patterns of microarchitectural modules, our mechanism can effectively limit simultaneous switching activity of close-by modules, thereby leveling voltage ringing at local power-pins. Compared to prior art, our di/dt controller is the first that takes the processor's floorplan as well as its power-pin distribution into account to provide a finer-grained control with minimal performance degradation. Based on the evaluation results using 2D floorplans, we show that our techniques can significantly improve inductive noise induced by current demand variation and reduce the average current variability by up to 7 times with an average performance overhead of 4.0%

Dataflow architectures provide an abundance of computing units that can be statically or dynamically configured to match the computing requirements of the given application. Wire delay has a reduced impact in dataflow architectures because only neighboring architectural entities are allowed to communicate within a single clock cycle. In this paper, the authors propose integer linear programming (ILP)-based placement and routing algorithms for mapping dataflow graphs (DFGs) to dataflow machines. The optimization process is guided by profiling information available from the compiler. The goal is to minimize the total execution time of the given application represented by a DFG under architectural constraints. A hierarchical method to handle the complexity of the initial ILP formulation is proposed. The profile-driven ILP algorithm reduces the total execution time of benchmark applications compared to the conventional wirelength-driven ILP approach. In addition, the ILP-based approach outperforms simulated annealing-based approach

As very large scale integration (VLSI) process technology migrates to nanoscale with a feature size of less than 100 nm, global wire delay is becoming a major hindrance in keeping the latency of intrachip communication within a single cycle, thus substantially decaying performance scalability. In addition, an effective microarchitectural floor planning algorithm can no longer ignore the dynamic communication patterns of applications. This article, using the profile information acquired at the microarchitecture level, proposes a "profile-guided microarchitectural floor planner" that considers both the impact of wire delay and the architectural behavior, namely, the intermodule communication, to reduce the latency of frequent routes inside a processor and to maintain performance scalability. Based on the simulation results here, the proposed profile-guided method shows a 5%-40% average instructions per cycle (IPC) improvement when the clock frequency is fixed. From the perspective of instruction throughput in billion instructions per second (BIPS), the floor planner is much more scalable than the conventional wirelength-based floor planner.

Three-dimensional (3-D) packaging via system-on-a-package (SOP) has been recently proposed as an alternative solution to overcome the limitation of system-on-a-chip (SOC) and meet the rigorous requirements of today's mixed signal system integration. The true potential of SOP technology lies in its capability to integrate both active and passive components into a single high speed/density 3-D packaging substrate. The routing environment for 3-D SOP is more advanced than that of the conventional printed circuit board (PCB) or multichip module (MCM) technology - pins are located at all layers of the SOP packaging substrate rather than the topmost layer only, and various types of vias are available for layer-to-layer connections. The contribution of this work is to provide: 1) the formulation of the new block-level 3-D global routing problem under wire length, layer, crosstalk, and congestion minimization and 2) the first global router for 3-D SOP named 3PGR. This paper reviews various approaches for MCM routing algorithms and investigates their applicability to the SOP model. The related experimental results demonstrate the effectiveness of the algorithms

As the size and complexity of very large scale integrated (VLSI) circuits increase, the need for faster floorplanning algorithms also grows. This paper introduces trapezoidal floorplanning for integrated circuits (Traffic), a new method for creating wire- and area-optimized floorplans. Through the use of connectivity grouping, simple geometry, and a constrained brute-force approach, Traffic achieves an average of 18% lower wire estimate than simulated annealing (SA) in orders of magnitude less time. This speed allows designers to rapidly explore a large circuit design space, to evaluate small changes to big circuits, to fit bounding boxes, and to produce initial solutions for other floorplanning algorithms.

Modern advances in reconfigurable analog technologies are allowing field-programmable analog arrays (FPAAs) to dramatically grow in size, flexibility, and usefulness. Our goal in this paper is to develop the first placement algorithm for large-scale floating-gate-based FPAAs with a focus on the minimization of the parasitic effects on interconnects under various device-related constraints. Our FPAA clustering algorithm first groups analog components into a set of clusters so that the total number of routing switches used is minimized and all IO paths are balanced in terms of routing switches used. Our FPAA placement algorithm then maps each cluster to a computational analog block (CAB) of the target FPAA while focusing on routing switch usage and balance again. Experimental results demonstrate the effectiveness of our approach

System-on-package (SOP) is a viable alternative to system-on-chip (SOC) for meeting the rigorous requirements of today's mixed-signal system integration. Thermal integrity is arguably the most crucial issue in three-dimensional (3-D) SOP due to the compact nature of the 3-D integration. In addition, the power supply noise issue becomes more serious as the supply voltage continues to decrease while the number of active devices consuming power increases. We propose a 3-D module and decap (decoupling capacitance) placement algorithm that evenly distributes the thermal profile and reduces the power supply noise. In addition, we allocate white spaces around the modules that require decaps to suppress the power supply noise while minimizing the area overhead. In our experimentation, we achieve improvements in both maximum temperature and decap amount with only small increase in area, wirelength, and runtime.

The recent popularity of 3D IC technology stems from its enhanced performance capabilities and reduced wirelength. However, wire congestion and thermal issues are exacerbated due to the compact nature of these layered technologies. In this paper, we develop techniques to reduce the maximum temperature and wire congestion of 3D circuits without compromising total wirelength and via count. Our approach consists of two phases. First, we use a multi-level min-cut placement with a modified gain function for local wire congestion and dynamic power consumption reduction. Second, we perform simulated annealing together with full-length thermal analysis and global routing for global wire congestion and maximum temperature reduction. Our experimental results show smooth tradeoff among congestion, temperature, wirelength, and via.

As wire delay increasingly becomes a significant performance bottleneck in monolithic architectures, there is a strong motivation to move to Dataflow Architectures. In this paper, we propose a set of placement algorithms for generic dataflow architectures. Our timing-driven and profile-driven placement algorithms respectively are targeting streaming and non-streaming applications. Compared to the conventional wirelength-driven algorithm, our timing-driven placer reduces the longest path delay by 23% and maximum slack by 26% at the cost of 10% increase in wirelength for streaming applications. In addition, our profile-driven placer reduces the total execution time of non-streaming applications by 17%. Lastly, our simultaneous timing/profile-driven placer reduces the total execution time of non-streaming applications by 13% on average.

Quantum-dot Cellular Automata (QCA) is a novel computing mechanism that can represent binary information based on spatial distribution of electron charge configuration in chemical molecules. In this paper, we present partitioning and placement algorithms for a large-scale automatic QCA layout. The purpose of zone partitioning is to initially partition a given circuit such that a single clock potential modulates the interdot barriers in all of the QCA cells within each zone. We then place these zones during our placement step. We identify several objectives and constraints that will enhance the buildability of QCA circuits and use them in our optimization process. The results are intended to define what is computationally interesting and could actually be built within a set of predefined constraints.

Modern advances in reconfigurable analog technologies are allowing field-programmable analog arrays (FPAAs) to dramatically grow in size, flexibility, and usefulness. Our goal in this paper is to develop the first hierarchical placement algorithm for large-scale floating-gate based FPAAs with a focus on the minimization of the parasitic effects on interconnects under various device-related constraints. In our hierarchical approach, our FPAA clustering algorithm first groups analog components into a set of clusters so that the total number of routing switches used is minimized and all IO paths are balanced in terms of routing switches used. Our FPAA placement algorithm then maps each cluster to a computational analog block (CAB) of the target FPAA while focusing on routing switch usage again. Experimental results demonstrate the effectiveness of our approach.

Dataflow architectures provide an abundance of computing units that can be statically or dynamically configured to match the computing requirements of the given application. Wire delay has a reduced impact in dataflow architectures because only neighboring architectural entities are allowed to communicate within a single clock cycle. In this paper, we propose MILP-based placement and routing algorithms for mapping dataflow graphs to dataflow machines. The optimization process is guided by profiling information available from the compiler. Our goal is to minimize the total execution time of the given application represented by a dataflow graph under architectural constraints. We propose a hierarchical method to handle the complexity of the initial MILP formulation. Our profile-driven MILP algorithm reduces the total execution time of benchmark applications compared to the conventional wirelength-driven approach on average by 18%.

3D packaging via System-On-Package (SOP) is a viable alternative to System-On-Chip (SOC) to meet the rigorous requirements of today's mixed signal system integration. In this work, we propose a 3D module and decap (decoupling capacitance) placement algorithm that simultaneously reduces the power supply noise and wire congestion. We provide efficient algorithms for 3D power supply noise and congestion analysis to guide our 3D module placement process. In addition, we allocate white spaces around the modules that require decaps to suppress the power supply noise while minimizing the area overhead. In our experimentation, we achieve improvements in both decap amount and congestion with only small increase in area, wirelength, and runtime.

Quantum-dot Cellular Automata (QCA) is a novel computing mechanism that can represent binary information based on spatial distribution of electron charge configuration in chemical molecules. QCA layout is currently restricted to a single layer with very limited number of wire crossing permitted. Thus, wire crossing minimization is crucial in improving the manufacturability of QCA circuits. In this article, we present the first QCA channel routing algorithm for wire crossing minimization. Our channel routing algorithm is able to reduce crossings, where Left-Edge First, Yoshimura and Kuh, and topologically-based algorithms fail to do so.

Modern advances in reconfigurable analog technologies are allowing field-programmable analog arrays (FPAAs) to dramatically grow in size, flexibility, and usefulness. With these advances, analog circuits and systems can be programmable, reconfigurable, adaptive, implemented on standard CMOS to take advantage of scaled CMOS technology, and at a density comparable to digital memories. Our goal in this paper is to develop the first physical design automation toolset for floating-gate based FPAA with focus on minimization of parasitic effects on FPAA interconnect. We provide graph-based analog circuit and FPAA device modeling suitable for efficient mapping. Our FPAA clustering algorithm constructs Computational Analog Blocks (CAB) from analog circuit elements while improving the utilization of the device and reducing its impact on the total number of routing switches used. Experimental results demonstrate the effectiveness of our approach.

Quantum-dot Cellular Automata (QCA) is a novel computing mechanism that can represent binary information based on spatial distribution of an electron charge configuration in chemical molecules. In this article, we present the first partitioning and placement algorithm for automatic QCA layout. We identify several objectives and constraints that will enhance the buildability of QCA circuits. The results are intended to: (1) define what is computationally interesting and could actually be built within a set of predefined constraints, (2) project what designs will be possible as additional constructs become realizable, and (3) provide a vehicle that we can use to compare QCA systems to silicon-based systems.

As process technology migrates to deep submicron with feature size less than 100nm, global wire delay is becoming a major hindrance in keeping the latency of intra-chip communication within a single cycle, thus decaying the performance scalability substantially. An effective floorplanning algorithm can no longer ignore the information of dynamic communication patterns of applications. In this paper, using the profile information acquired at the architecture/microarchitecture level, we propose a "profile-guided microarchitectural floorplanner" that considers both the impact of wire delay and the architectural behavior, namely the inter-module communication, to reduce the latency of frequent routes inside a processor and to maintain performance scalability. Based on our simulation results, the profile-guided method shows a 5% to 40% average IPC improvement when clock frequency is fixed. From the perspective of instruction throughput(in BIPS), our floorplanner is much more scalable than a conventional wire length based floorplanner.

In this paper, we study the net and pin distribution problem for global routing targeting three dimensional packaging layout via system-on-package (SOP). The routing environment for the new emerging mixed-signal SOP technology is more advanced than that of the conventional PCB or MCM technology - pins are located at all layers of SOP packaging substrate rather than the top-most layer only. This is the first work to formulate and solve the multi-layer net and pin distribution for layer, wirelength, and crosstalk minimization.

Quantum-dot Cellular Automata (QCA) is a novel computing mechanism that can represent binary information based on spatial distribution of electron charge configuration in chemical molecules. It has the potential to allow for circuits and systems with functional densities that are better than end of the roadmap CMOS, but also imposes new constraints on system designers. In this paper we develop the first cell-level placement of QCA circuits, where the given circuit is assumed to be partitioned into 4-phase asynchronous QCA timing zones. We formulate the QCA cell placement in each timing zone as a unidirectional geometric embedding of k-layered bipartite graphs. We then present an analytical and a stochastic solution for minimizing the wire crossings and wire length in these placement solutions.

In this paper, we propose a new efficient O(nlogn) connectivity-based bottom-up clustering algorithm called edge separability-based clustering (ESC). Unlike existing bottom-up algorithms that are based on local connectivity information of the netlist, ESC exploits more global connectivity information using edge separability to guide the clustering process, while carefully monitoring cluster area balance. Exact computation of the edge separability /spl lambda/(e) for a given edge e=(x,y) in an edge-weighted undirected graph G is equivalent to finding the maximum flow between x and y. Since the currently best known time bounds for solving the maximum flow problem is O(mnlog(n/sup 2//m)), due to Goldberg and Tarjan (Goldberg and Tarjan, 1988), the computation of /spl lambda/(e) for all edges in G requires O(m/sup 2/nlog(n/sup 2//m)) time. However, we show that a simple and efficient algorithm CAPFOREST (Nagamochi and Ibaraki, 1992) can be used to provide a good approximation of edge separability (within 9.1% empirical error bound) for all edges in G without using any network flow computation in O(nlogn) time. Our experimental results based on large-scale benchmark circuits demonstrate the effectiveness of using edge separability in the context of multilevel partitioning framework for cutsize minimization. We observe that exploiting edge separability yields better quality partitioning solution compared to existing clustering algorithms (Sun and Sechen, 1993), (Cong and Smith, 1993), (Huang and Kahng, 1995), (Ng et al., 1987), (Wei and Cheng, 1991), (Shin and Kim, 1993), (Schuler and Ulrich, 1972), (Karypis et al., 1997), that rely on local connectivity information. In addition, our ESC-based iterative improvement based multilevel partitioning algorithm LR/ESC-PM provides comparable results to state-of-the-art hMetis package (Karypis et al., 1997), (Karypis and Kumar, 1999).

In this paper, we formulate the physical planning with retiming problem and propose an algorithm called GEO. Our performance-driven global placement algorithm GEO is mincut-based, where a multilevel partitioning is performed recursively to divide the netlist and assign gates to the tiles in a top-down fashion. The contribution of our work is on the development of retiming-aware timing analysis (RTA) that is used to guide our mincut-based global placement. Compared to the conventional static timing analysis, RTA provides timing slack information after retiming so that the clock period after retiming can be directly minimized during the placement. We show how to make an effective use of RTA timing slack information in a multilevel partitioning framework. Simultaneous consideration of partitioning and retiming under the geometric delay model enables GEO to hide global interconnect latency more effectively compared to the conventional approaches. In our comparison to the state-of-the art methods that perform partitioning, retiming, and simulated annealing-based floorplanning separately, GEO obtains significant improvement on retimed delay, while maintaining comparable wirelength and runtime results.

Abstract| In this paper, w ein troduce a new efcient O(n log n) graph search based bottom-up clustering algorithm named ESC (Edge Separability based Clustering).Unlike existing bottom-up algorithms that are based on local connectivity information of the netlist, ESC exploits more global connectivity information \edge separability" to guide clustering process while carefully monitoring cluster area balance.Computing the edge separability for a given edge e = ( x; y) in an edge weighted undirected graph G(V; E; s; w) i s equivalent to nding the x-y mincut.Then, we show that a simple and ecient algorithm CAPFOREST [14] can be used to provide a good estimation of edge separability for all edges in G without using any netw ork o w computation.Related experiments based on large scale ISPD98 [1] benchmark circuits con rm that exploiting edge separability yields better quality partitioning solution compared to various bottom-up clustering algorithms proposed in the literature including Absorption [18], Density [9, 11], Rent P arameter [15], Ratio Cut [19], Closeness [17], and Connectivity [16] method.In addition, our ESC based multiway partitioning algorithm LR/ESC-PM pro vides comparable results to state-of-the-art hMetis [12] and hMetis-Kway [13].

Abstract | Under the interconnect-centric design paradigm, partitioning is seen as the crucial step that de nes the interconnect [1]. To meet the performance requirement of today's complex design, performance driven partitioners must consider the amount of interconnect induced by partitioning as well as its impact on performance. In this paper, we provide new performance driven formulation for cell move based top-down multiway partitioning algorithms with consideration of the local and global interconnect delay. In our \\constrained acyclic partitioning " formulation, cell moves are restricted to maintain acyclicity in partitioning solution to prevent cyclic dependency among cells in di erent partitions. In our \\relaxed acyclic partitioning " formulation, acyclic constraints are relaxed to give partitioners capability of minimizing cutsize and delay. Our new acyclic constraint based performance driven multiway partitioning algorithm FLARE obtains (i) 4 % to 13 % better delay compared to the state-of-the-art cutsize minimization based hMetis [10] at almost no increase in cutsize, and (ii) 84 % better cutsize compared to the state-of-theart delay minimization based PRIME [2] at an expense of 16 % increase in delay. I.

Abstract | The multi-way partition problem is very important in various applications. In this paper, we use analytical and experimental results to study the k-way partition problem. We introduce the concept of embedding graph for the the k-way partition problem. Based on this concept, we explain di erent scenarios of using a bi-partition heuristic to solve the k-way partition problem. If C denote the optimal cut cost for the k-way partition problem and the bi-partition heuristics we use are-approzimation heuristics (dened in Section 2), we prove that the cut cost from the hierarchical approach has an approximate upper bound of C log k while the cut cost from the all-way bipartition, or at approach, has an upper bound of Ck. This is contrary to some claims made in recent literature ( and CAD tools designed based on it). Experimental results strongly support our theoretical analysis. Our results show that for large target graph, the hierarchical approach is about 77 % better than the single-pass all-way bi-partition approach. The all-way bi-partition approach will perform better in a multipass set-up. However, the hierarchical approach is still on average 7:1 % better in quality and 144 times faster than the multi-partition all-way bi-partition approach. The main conclusion of this paper is, contradicted to what has been suggested in literature, hierarchical bipartitioning is a more e ective multi-way partitioning scheme. I.

In this paper, we study the performance driven multiw ay circuit partitioning problem with consideration of the significant difference of local and global interconnect delay induced by the partitioning. We develop an efficient algorithm HPM (Hierarc hicalP erformance driven Multi-level partitioning) that simultaneously considers cutsize and delay minimization with retiming. HPM builds a multi-lev el cluster hierarc hy and performs various refinement while gradually decomposing the clusters for simultaneous cutsize and delay minimization. We provide comprehensive experimental justification for each step involv ed in HPM and in-depth analysis of cutsize and delay tradeoff existing in the performance driven partitioning problem. HPM obtains (i) 7% to 23% better delay compared to the state-of-the-art cutsize driven hMetis [11] at the expense of 19% increase in cutsize, and (ii) 81% better cutsize compared to the state-of-the-art delay driven PRIME [2] at the expense of 6% increase in delay.

In this paper, we propose a unified approach to partitioning, floorplanning, and retiming for effective and efficient performance optimization. The integration enables the partitioner to exploit more realistic geometric delay model provided by the underlying floorplan. Simultaneous consideration of partitioning and retiming under the geometric delay model enables us to hide global interconnect latency effectively by repositioning FF along long wires. Under the proposed geometric embedding based performance driven partitioning problem, our GEO algorithm performs multi-level top-down partitioning while determining the location of the partitions. We adopt the concept of sequential arrival time and develop sequential required time in our retiming based timing analysis engine. GEO performs cluster-move based iterative improvement on top of multi-level cluster hierarchy, where the gain function obtained from the timing analysis is based on the minimization of cutsize, wirelength, and sequential slack. In our comparison to (i) state-of-the-art partitioner hMetis followed by retiming and simulated annealing based slicing floorplanning, and (ii) state-of-the-art simultaneous partitioning with retiming HPM followed by floorplanning, GEO obtains 35% and 23% better delay results while maintaining comparable cutsize, wirelength, and runtime results.

It is know to many researchers in the partitioning community that the recursive bipartitioning approach outperforms the direct non-recursive approach in solving the multiway partitioning problem.However, fittle progress has been made to identify and overcome the weakness of the direct (dtematively called flat) approach.In this paper, we make the fit observation that the performmce of iterative improvement-based flat multimy petitioner K-FM [10,11] is not suitable for today's large scale circuits.Then, we pr~ pose a simple yet effective hill-cfimbing method called PM (P-e cell hfovement) that overcomes the ~iitation of K-FAI and providw partitioners the capability to e\Tlore tider range of solution space effectively while ensuring convergence to satisfying suboptirnd solutions.The main idea is to reduce the mtitiway partitioning problem to sets of concurrent bipartitioning problems.Starting tith an initial mtitimy partition of the netfist, we apply 2-way FM [7] to pairs of blocks so as to improve the qufllty of overall multiwy partitioning solution.The pairiig of blocks is based on the gain of the last pass, and the Ptise cell Movement (Phi) passes continue until no further gain can be obtained.We observe that Phl passes are effective in distributing clusters evenly into mtitiple blocks to minimize the connections across the multiway cut~iea.Our iterative improvementbased flat multiway partitioned K-PM/LR improves K-FM by a surprising average mmgin of up to 86.2% and outperforms its counterpart recursive FM by up to 17.3% when tested on LICNC and large scale ISPD98 benchmark circuits [1].

In this paper, we present an efficient Iterative Improvement based Partitioning (IIP) algorithm called LSR/MFFS, that combines signal flow based Maximum Fanout Free Subgraph (MFFS) clustering algorithm with Loose and Stable net Removal (LSR) partitioning algorithm. The MFFS algorithm generalizes existing MFFC decomposition method from combinational circuits to general sequential circuits in order to handle cycles naturally. We also study the properties of the nets that straddle the cutline carefully, and introduce the concepts of the loose and stable nets as well as effective ways to remove them out of the cutset. The LSR/MFFS algorithm first applies LSR algorithm to clustered netlist generated by MFFS algorithm for global-level cutsize optimization and then declusters netlist for further cutsize refinement. As a result, the LSR/MFFS algorithm has achieved the best cutsize result among all the bipartitioning algorithms published in the literatures with very promising runtime performance. In particular, it outperforms the recent state-of-the-art IIP algorithms LA3-CDIP, CLIP-PROP/sub f/, Strawman, hMetis-FM, and MLc by 17.4%, 12.1%, 5.9%, 3.1%, and 1.9%, respectively. It also outperforms the state-of-the-art non-IIP algorithms Paraboli, FEB, and PANZA by 32.0%, 21.4%, and 1.4%, respectively.