Detailed routing remains one of the most complex and time-consuming steps in modern physical design due to the challenges posed by shrinking feature sizes and stricter design rules. Prior detailed routers achieve state-of-the-art results by leveraging iterative pathfinding algorithms to route each net. However, runtimes are a major issue in detailed routers, as converging to a solution with zero design rule violations (DRVs) can be prohibitively expensive.In this paper, we propose leveraging reinforcement learning (RL) to enable rapid convergence in detailed routing by learning from previous designs. We make the key observation that prior detailed routers statically schedule the cost weights used in their routing algorithms, meaning they do not change in response to the design or technology. By training a conservative Q-learning (CQL) model to dynamically select the routing cost weights to minimize the number of algorithm iterations, we find that our work completes the ISPD19 benchmarks with 1.56 × average and up to 3.01 × faster runtime than the baseline router while maintaining or improving the DRV count in all cases. We also find that this learning shows signs of generalization across technologies, meaning that learning designs in one technology can translate to improved outcomes in other technologies.

Tightly integrated optical interconnects can provide high-bandwidth, energy-efficient inter-node communication. We describe a novel system which uses hybrid 2.5D/3D integration to compose a state-of-the-art FPGA compute chiplet, three electrical interface chiplets, and three photonic interface chiplets. We use register-transfer-, gate-, transistor-, and device-level simulations to demonstrate the potential for this system to achieve 96Tb/s of bi-directional bandwidth, and we experimentally demonstrate key components including a complete opto-electrical channel. Our results provide a strong case for hybrid 2.5D/3D integration as the key enabler for scaling co-packaged optical interconnects.

The electronic design automation (EDA) community has recently begun recognizing the potential of generative artificial intelligence (AI) in chip design. However, its full potential is not fully exploited due to the limited availability of publicly accessible datasets crucial for advancing research in EDA. This paper highlights the dual role of generative AI; in particular, it showcases (i) BeGAN, the use of a generative AI strategy to create thousands of realistic benchmarks for power grid synthesis and analysis to advance power-related research, and (ii) EDA Corpus---an expert-curated and generative AI-enhanced dataset to serve research and development of EDA tool assistants. These two case studies emphasize the ability of generative methods to create and utilize datasets to advance research and lower the barriers to entry in EDA.

Several academic EDA tools have been released; however, few are used in real tapeouts, even by other academics. Robust open-source tools require feedback and direction from users. To this end, OpenROAD employs end-users as internal design advisors who bring with them the experience of multiple tapeouts and EDA tool flow development. This paper discusses the OpenROAD design advisors' ongoing work to bring OpenROAD from a collection of tools to an end-to-end autonomous design flow. We discuss our work to fill in the gaps for a full RTL-to-GDS design flow, assemble a full-flow test suite reflective of real tapeouts, debug flow-level issues between tools, and bridge the gap between OpenROAD developers and others in the open-source community. Lastly, we discuss OpenROAD's long-term goal to become fully autonomous, and what that means from a user's perspective.

While systolic arrays are widely used for dense-matrix operations, they are seldom used for sparse-matrix operations. In this paper, we show how a systolic array of Multiply-and-Accumulate (MAC) units, similar to Google's Tensor Processing Unit (TPU), can be adapted to efficiently handle sparse matrices. TPU-like accelerators are built upon a 2D array of MAC units and have demonstrated high throughput and efficiency for dense matrix multiplication, which is a key kernel in machine learning algorithms and is the target of the TPU. In this work, we employ a co-designed approach of first developing a packing technique to condense a sparse matrix and then propose a systolic array based system, Sparse-TPU, abbreviated to STPU, to accommodate the matrix computations for the packed denser matrix counterparts. To demonstrate the efficacy of our co-designed approach, we evaluate sparse matrix-vector multiplication on a broad set of synthetic and real-world sparse matrices. Experimental results show that STPU delivers 16.08X higher performance while consuming 4.39X and 19.79X lower energy for integer (int8) and floating point (float32) implementations, respectively, over a TPU baseline. Meanwhile, STPU has 12.93% area overhead and an average of 4.14% increase in dynamic energy over the TPU baseline for the float32 implementation.

The computation for today's intelligent personal assistants such as Apple Siri, Google Now, and Microsoft Cortana, is performed in the cloud. This cloud-only approach requires significant amounts of data to be sent to the cloud over the wireless network and puts significant computational pressure on the datacenter. However, as the computational resources in mobile devices become more powerful and energy efficient, questions arise as to whether this cloud-only processing is desirable moving forward, and what are the implications of pushing some or all of this compute to the mobile devices on the edge.

With silicon-based transistors approaching their scaling limits, multiple successor technologies are competing for silicon's place. Due to recent fabrication breakthroughs, one promising alternative is the carbon nanotube field-effect transistor (CNTFET), which uses carbon nanotubes as the channel medium instead of silicon. Although logic gates using CNTFETs have been demonstrated to provide up to an order of magnitude better energy-delay product (EDP) over silicon-based counterparts, system-level design using CNTFETs show significantly smaller EDP improvement because of the critical path of the design, output load capacitance and corresponding drive strengths of gates. In this paper, we address this challenge by exploring various architectural design choices using CNTFET-based pass transistor logic (PTL) and create an energy-efficient RISC-V processor. While silicon-based design traditionally prefers complementary logic over PTL, CNTFETs are ideal candidates for PTL due to their low threshold voltage, low power dissipation, and equal strength p-type and n-type transistors. By utilizing PTL to design modules that lie on the processor's critical path, systems can efficiently exploit CNTFET's potential benefits. Our results show that while a CNTFET RISC-V processor using complementary logic achieves a 2.9× EDP improvement over a silicon design, using PTL along the critical path components in the ALU can boost EDP improvement 5× as well as reduce area by 17% over 16 nm silicon CMOS.

As user demand scales for intelligent personal assistants (IPAs) such as Apple’s Siri, Google’s Google Now, and Microsoft’s Cortana, we are approaching the computational limits of current datacenter (DC) architectures. It is an open question how future server architectures should evolve to enable this emerging class of applications, and the lack of an open-source IPA workload is an obstacle in addressing this question. In this article, we present the design of Sirius, an open end-to-end IPA Web-service application that accepts queries in the form of voice and images, and responds with natural language. We then use this workload to investigate the implications of four points in the design space of future accelerator-based server architectures spanning traditional CPUs, GPUs, manycore throughput co-processors, and FPGAs. To investigate future server designs for Sirius, we decompose Sirius into a suite of eight benchmarks (Sirius Suite) comprising the computationally intensive bottlenecks of Sirius. We port Sirius Suite to a spectrum of accelerator platforms and use the performance and power trade-offs across these platforms to perform a total cost of ownership (TCO) analysis of various server design points. In our study, we find that accelerators are critical for the future scalability of IPA services. Our results show that GPU- and FPGA-accelerated servers improve the query latency on average by 8.5× and 15×, respectively. For a given throughput, GPU- and FPGA-accelerated servers can reduce the TCO of DCs by 2.3× and 1.3×, respectively.

As user demand scales for intelligent personal assistants (IPAs) such as Apple's Siri, Google's Google Now, and Microsoft's Cortana, we are approaching the computational limits of current datacenter architectures. It is an open question how future server architectures should evolve to enable this emerging class of applications, and the lack of an open-source IPA workload is an obstacle in addressing this question. In this paper, we present the design of Sirius, an open end-to-end IPA web-service application that accepts queries in the form of voice and images, and responds with natural language. We then use this workload to investigate the implications of four points in the design space of future accelerator-based server architectures spanning traditional CPUs, GPUs, manycore throughput co-processors, and FPGAs.