In this paper, we address three core problems in the layout synthesis of multi-row standard cells: transistor folding, row partitioning, and transistor placement. We propose a comprehensive solution to the problem of synthesizing area-optimal multi-row standard cells by seamlessly integrating transistor folding and row partitioning into the transistor placement framework. Additionally, we introduce a systematic methodology to construct a standard cell library. This methodology determines the cell types among single-row, multi-row with VDD-abut, and multi-row with VSS-abut in order to achieve an optimal trade-off between power, performance, and area (PPA) for the target design implementation. Experimental results demonstrate that for 4-routing track standard cells of the advanced technology nodes our multi-row cell generator increases the cell generation completion ratio from 72% to 100% and reduces the metal length for in-cell routing by 11.9% while maintaining comparable cell area compared to the area-minimal single-row cells. Furthermore, using our optimized cell library is able to reduce the target chip area by 4.2% and chip power by 7.0%, while all meeting timing and design rule constraints, compared to the state-of-the-art single-row standard cell library.

Resolving the design rule checking (DRC) violations at the pre-route stage is critically important to reduce the time-consuming design closure process at the post-route stage. Recently, noticeable methodologies have been proposed to predict DRC hotspots using Machine Learning based prediction models. However, little attention has been paid to how the predicted DRC violations can be effectively resolved. In this paper, we propose a pre-route DRC violation resolution methodology that is tightly coupled with fully compatible prediction model. Precisely, we devise different resolution strategies for two types of DRC violations: (1) pin accessibility (PA)-related and (2) routing congestion (RC)-related. To this end, we develop a fully predictable ML-based model for both PA and RC-related DRC violations, and propose completely different resolution techniques to be applied depending on the DRC violation type informed by the compatible prediction model such that for (1) PA-related DRC violation, we extract the DRC violation mitigating regions, then improve placement by formulating the whitespace redistribution problem on the regions into an instance of Bayesian Optimization problem to produce an optimal cell perturbation, while for (2) RC-related DRC violation, we manipulate the routing resources within the regions that have high potential for the occurrence of RC-related DRC violation. Through experiments, it is shown that our methodology is able to resolve the number of DRC violations by 26.54%, 25.28%, and 20.34% further on average over that by a conventional flow with no resolution, a commercial ECO router, and a state-of-the-art academic predictor/resolver, respectively, while maintaining comparable design quality.

Deep learning (DL) models have recently paid considerable attention to timing prediction in the place-and-route (P&R) flow. As yet, the DL-based prior works are confined to timing prediction at the time-consuming routing stage, and very few have addressed the timing prediction problem at the placement, i.e., at the pre-route stage. Moreover, no work has addressed a seamless link of timing prediction at the pre-route stage to the final timing optimization through commercial P&R tools. In this work, we introduce a novel framework called DTOC-P that seamlessly integrates deep-learning-driven timing optimization into cutting-edge commercial P&R tools. Our framework is composed of two phases: (1) the pre-route timing prediction phase that performs DL-driven arc delay and arc output slew prediction with an elaborated hierarchical model; (2) the timing optimization phase which incorporates commercial P&R tools with DL-driven prediction outcomes to perform timing optimization. In addition, DTOC-P framework achieves enhanced practicality with the application of continual learning in the timing prediction phase, and the concept of anomaly detection in the timing optimization phase. Experimental results show that our DTOC-P framework improves pre-route prediction accuracy by up to 55% and 47% on arc delay and arc output, which are further enhanced to encompass a broader range of designs by continual learning supported in DTOC-P, practically using a tenfold reduced training time compared to re-training all datasets from scratch. In terms of timing optimization, our experiments reveal that DTOC-P framework improves WNS, TNS, and the number of timing violation paths by up to 12%, 41%, and 34%, respectively, which is a remarkable progress compared to its predecessor through the integration of anomaly detection that excludes potential outliers to effectively protects against erroneous timing updates during the timing optimization phase.

Advanced technology nodes face challenges related to DRVs (design rule violations), primarily due to (1) pin inaccessibility and routing on congested region. While various ML (machine learning) techniques have been introduced to address these issues during placement, aggregating data on pin accessibility and routing congestion for ML model training has proven very challenging. This study presents an innovative ML-based approach to DRC (design rule check) hotspot prediction that effectively captures the combined impact of pin accessibility and routing congestion. Specifically, we introduce the concept of pin proximity graph, which accurately represents spatial information regarding cell I/O pins and pin-to-pin disturbance relationships. We then propose a novel ML model called PGNN, which seamlessly integrates GNN (Graph Neural Network) and U-net. In this approach, GNN handles the incorporation of pin accessibility information derived from the pin proximity graph while U-net extracts routing congestion information from grid-based features. Additionally, we solidify the capability of our prediction model toward ensuring the practical applicability and sustainability of our model by integrating two learning methodologies into our model training framework. Those are (1) transfer learning whose objective is to retain the same level of prediction accuracy in spite of not having enough data on the new process node and (2) incremental learning whose objective is to reduce the train time while maintaining the model accuracy in similar quality when new circuits are added.

An accurate DRC (design rule check) hotspot prediction at the placement stage is essential in order to reduce a substantial amount of design time required for the iterations of placement and routing. It is known that for implementing chips with advanced technology nodes, (1) pin accessibility and (2) routing congestion are two major causes of DRVs (design rule violations). Though many ML (machine learning) techniques have been proposed to address this prediction problem, it was not easy to assemble the aggregate data on items 1 and 2 in a unified fashion for training ML models, resulting in a considerable accuracy loss in DRC hotspot prediction. This work overcomes this limitation by proposing a novel ML based DRC hotspot prediction technique, which is able to accurately capture the combined impact of items 1 and 2 on DRC hotspots. Precisely, we devise a graph, called pin proximity graph, that effectively models the spatial information on cell I/O pins and the information on pin-to-pin disturbance relation. Then, we propose a new ML model, called PGNN, which tightly combines GNN (graph neural network) and U-net in a way that GNN is used to embed pin accessibility information abstracted from our pin proximity graph while U-net is used to extract routing congestion information from grid-based features. Through experiments with a set of benchmark designs using Nangate 15nm library, our PGNN outperforms the existing ML models on all benchmark designs, achieving on average 7.8~12.5% improvements on F1-score while taking 5.5× fast inference time in comparison with that of the state-of-the-art techniques.

Near-threshold computing is essential for energy-efficient operation of VLSI systems, but wide performance variation and nonlinearity to process variations block the proliferation. To cope with this, in this article, we propose a holistic hardware performance monitoring methodology for accurate timing prediction in a near-threshold voltage regime. Precisely, 1) we formulate the problem of finding an efficient configuration of monitoring circuits into an instance of optimal experiment design problem and 2) propose a new timing prediction flow, consisting of statistical estimation of FEOL and BEOL process variations and a neural network-based timing inference model. For accurate control of timing margin and overcoming simulation-silicon discrepancies, 3) we introduce uncertainty learning in the prediction model construction and calibrate it through transfer learning. Furthermore, 4) we avoid time-consuming SPICE simulations in our methodology by employing efficient but accurate surrogate models. Through simulations using a 28-nm industry PDK and DK characterized at 0.6-V operation, it is shown that our methodology is highly effective, reducing the average prediction pessimism of maximum delay by 77.9% over conventional signoff results while respecting target prediction yield. Besides, for test chips fabricated using a 10-nm process, we demonstrated that our holistic approach from design to postsilicon phase in conjunction with adaptive voltage scaling reduces dynamic power consumption by 28.2%–28.8% on average, in comparison with typical supply voltage operation.

Simulation-based functional verification is characterized by two inherently conflicting targets: the signal visibility and simulation performance. Achieving a proper trade-off between these two targets is of paramount importance. Even though HDL simulators are the most widely used verification platform at the RTL and gate level, their major drawback is the low performance in verifying complex SOCs, especially when the high visibility over the design under verification is required. This paper presents a new, fast simulation method as an effective way to achieve both high simulation speed and full signal visibility. It is based on an original two-pass simulation approach. During the 1stpass, with the simulation running at full speed, a set of design states is saved periodically at predetermined checkpoints. During the 2ndpass, another simulation is performed, using any of saved checkpoints and providing 100% signal visibility for debugging. Our method differs from the traditional simulation snapshot approach in the amount and the way the design state is saved. Experimental results show significant speed-up compared to existing traditional simulation methods while maintaining 100% visibility.

We present a robust datapath allocation method that is flexible enough to handle constraints imposed by a variety of target architectures. Key features of this method are its ability to handle accurate modeling of datapath units and the simultaneous optimization of direct objective functions. The proposed method consists of a new binding model construction scheme and an optimization technique based on simulated annealing. To illustrate the flexibility of this method, two datapath allocation procedures have been developed for two problem enviroments: (1) a procedure that incorporates interconnection area and delay estimates, where floor-planning is tightly integrated into datapath allocation; and (2) a procedure that handles registers, register files, and multiport memories for data storage, as well as random and linear topologies for interconnection architectures. Results from these two applications show our method produces competitive designs for benchmark circuits, as well as being flexible enough to be used for a variety of different domains.

We present a novel technique for datapath allocation, which incorporates interconnection area and delay estimates based on dynamic floorplanning. In this approach, datapath area is minimized by minimizing the number of wires, routing tracks, and multiplexers, while performance is optimized by minimizing wire length. The simultaneous optimization of these physical cost metrics allows the system to explore realistic design solutions.