arXiv subjects

Jason Cong

Publications and source records attributed to Jason Cong.

At least 73 records · Page 4Linked to original sources

arXiv2024

Cross-Modality Program Representation Learning for Electronic Design Automation with High-Level Synthesis

In recent years, domain-specific accelerators (DSAs) have gained popularity for applications such as deep learning and autonomous driving. To facilitate DSA designs, programmers use high-level synthesis (HLS) to compile a high-level description written in C/C++ into a design with low-level hardware description languages that eventually synthesize DSAs on circuits. However, creating a high-quality HLS design still demands significant domain knowledge, particularly in microarchitecture decisions expressed as \textit{pragmas}. Thus, it is desirable to automate such decisions with the help of machine learning for predicting the quality of HLS designs, requiring a deeper understanding of the program that consists of original code and pragmas. Naturally, these programs can be considered as sequence data. In addition, these programs can be compiled and converted into a control data flow graph (CDFG). But existing works either fail to leverage both modalities or combine the two in shallow or coarse ways. We propose ProgSG, a model that allows interaction between the source code sequence modality and the graph modality in a deep and fine-grained way. To alleviate the scarcity of labeled designs, a pre-training method is proposed based on a suite of compiler's data flow analysis tasks. Experimental results show that ProgSG reduces the RMSE of design performance predictions by up to $22\%$, and identifies designs with an average of $1.10\times$ and $1.26\times$ (up to $8.17\times$ and $13.31\times$) performance improvement in design space exploration (DSE) task compared to HARP and AutoDSE, respectively.

cs.LG↗

arXiv2024

Compiling Quantum Circuits for Dynamically Field-Programmable Neutral Atoms Array Processors

Dynamically field-programmable qubit arrays (DPQA) have recently emerged as a promising platform for quantum information processing. In DPQA, atomic qubits are selectively loaded into arrays of optical traps that can be reconfigured during the computation itself. Leveraging qubit transport and parallel, entangling quantum operations, different pairs of qubits, even those initially far away, can be entangled at different stages of the quantum program execution. Such reconfigurability and non-local connectivity present new challenges for compilation, especially in the layout synthesis step which places and routes the qubits and schedules the gates. In this paper, we consider a DPQA architecture that contains multiple arrays and supports 2D array movements, representing cutting-edge experimental platforms. Within this architecture, we discretize the state space and formulate layout synthesis as a satisfiability modulo theories problem, which can be solved by existing solvers optimally in terms of circuit depth. For a set of benchmark circuits generated by random graphs with complex connectivities, our compiler OLSQ-DPQA reduces the number of two-qubit entangling gates on small problem instances by 1.7x compared to optimal compilation results on a fixed planar architecture. To further improve scalability and practicality of the method, we introduce a greedy heuristic inspired by the iterative peeling approach in classical integrated circuit routing. Using a hybrid approach that combined the greedy and optimal methods, we demonstrate that our DPQA-based compiled circuits feature reduced scaling overhead compared to a grid fixed architecture, resulting in 5.1X less two-qubit gates for 90 qubit quantum circuits. These methods enable programmable, complex quantum circuits with neutral atom quantum computers, as well as informing both future compilers and future hardware choices.

quant-ph↗

arXiv2024

Depth-Optimal Addressing of 2D Qubit Array with 1D Controls Based on Exact Binary Matrix Factorization

Reducing control complexity is essential for achieving large-scale quantum computing. However, reducing control knobs may compromise the ability to independently address each qubit. Recent progress in neutral atom-based platforms suggests that rectangular (row-column) addressing may strike a balance between control granularity and flexibility for 2D qubit arrays. This scheme allows addressing qubits on the intersections of a set of rows and columns each time. While quadratically reducing controls, it may necessitate more depth. We formulate the depth-optimal rectangular addressing problem as exact binary matrix factorization, an NP-hard problem also appearing in communication complexity and combinatorial optimization. We introduce a satisfiability modulo theories-based solver for this problem, and a heuristic, row packing, performing close to the optimal solver on various benchmarks. Furthermore, we discuss rectangular addressing in the context of fault-tolerant quantum computing, leveraging a natural two-level structure.

cs.ET↗

arXiv2024

TAPA-CS: Enabling Scalable Accelerator Design on Distributed HBM-FPGAs

Despite the increasing adoption of Field-Programmable Gate Arrays (FPGAs) in compute clouds, there remains a significant gap in programming tools and abstractions which can leverage network-connected, cloud-scale, multi-die FPGAs to generate accelerators with high frequency and throughput. To this end, we propose TAPA-CS, a task-parallel dataflow programming framework which automatically partitions and compiles a large design across a cluster of FPGAs with no additional user effort while achieving high frequency and throughput. TAPA-CS has three main contributions. First, it is an open-source framework which allows users to leverage virtually "unlimited" accelerator fabric, high-bandwidth memory (HBM), and on-chip memory, by abstracting away the underlying hardware. This reduces the user's programming burden to a logical one, enabling software developers and researchers with limited FPGA domain knowledge to deploy larger designs than possible earlier. Second, given as input a large design, TAPA-CS automatically partitions the design to map to multiple FPGAs, while ensuring congestion control, resource balancing, and overlapping of communication and computation. Third, TAPA-CS couples coarse-grained floorplanning with automated interconnect pipelining at the inter- and intra-FPGA levels to ensure high frequency. We have tested TAPA-CS on our multi-FPGA testbed where the FPGAs communicate through a high-speed 100Gbps Ethernet infrastructure. We have evaluated the performance and scalability of our tool on designs, including systolic-array based convolutional neural networks (CNNs), graph processing workloads such as page rank, stencil applications like the Dilate kernel, and K-nearest neighbors (KNN). TAPA-CS has the potential to accelerate development of increasingly complex and large designs on the low power and reconfigurable FPGAs.

cs.DC↗

arXiv2024

Quantum State Preparation Using an Exact CNOT Synthesis Formulation

Minimizing the use of CNOT gates in quantum state preparation is a crucial step in quantum compilation, as they introduce coupling constraints and more noise than single-qubit gates. Reducing the number of CNOT gates can lead to more efficient and accurate quantum computations. However, the lack of compatibility to model superposition and entanglement challenges the scalability and optimality of CNOT optimization algorithms on classical computers. In this paper, we propose an effective state preparation algorithm using an exact CNOT synthesis formulation. Our method represents a milestone as the first design automation algorithm to surpass manual design, reducing the best CNOT numbers to prepare a Dicke state by 2x. For general states with up to 20 qubits, our method reduces the CNOT number by 9% and 32% for dense and sparse states, on average, compared to the latest algorithms.

cs.IT↗

arXiv2023

ProgSG: Cross-Modality Representation Learning for Programs in Electronic Design Automation

Recent years have witnessed the growing popularity of domain-specific accelerators (DSAs), such as Google's TPUs, for accelerating various applications such as deep learning, search, autonomous driving, etc. To facilitate DSA designs, high-level synthesis (HLS) is used, which allows a developer to compile a high-level description in the form of software code in C and C++ into a design in low-level hardware description languages (such as VHDL or Verilog) and eventually synthesized into a DSA on an ASIC (application-specific integrated circuit) or FPGA (field-programmable gate arrays). However, existing HLS tools still require microarchitecture decisions, expressed in terms of pragmas (such as directives for parallelization and pipelining). To enable more people to design DSAs, it is desirable to automate such decisions with the help of deep learning for predicting the quality of HLS designs. This requires us a deeper understanding of the program, which is a combination of original code and pragmas. Naturally, these programs can be considered as sequence data, for which large language models (LLM) can help. In addition, these programs can be compiled and converted into a control data flow graph (CDFG), and the compiler also provides fine-grained alignment between the code tokens and the CDFG nodes. However, existing works either fail to leverage both modalities or combine the two in shallow or coarse ways. We propose ProgSG allowing the source code sequence modality and the graph modalities to interact with each other in a deep and fine-grained way. To alleviate the scarcity of labeled designs, a pre-training method is proposed based on a suite of compiler's data flow analysis tasks. Experimental results on two benchmark datasets show the superiority of ProgSG over baseline methods that either only consider one modality or combine the two without utilizing the alignment information.

cs.LG↗

arXiv2023

Locality and Utilization in Placement Suboptimality

The mixed-size placement benchmarks described in this book chapter directly address several of the shortcomings in previously published suboptimality benchmarks. Two new sets of placement examples are constructed, one targeting the role of nonlocal nets in suboptimality, and another targeting the role of white space and large variations in module sizes. The first set, PEKO-MC, is a set of standard-cell circuits with nonlocal nets in known optimal placements. A given netlist is modified so as to render a given placement for it optimal for the new netlist. Cell dimensions and locations are not changed, net-degree statistics are matched exactly, and over 60% of the original netlist is left unchanged. The second set, PEKO-MS, incorporates a parametrized percentage of white space into a mixed-size placement which precisely matches given macro dimensions and locations as well as the net-degree distributions of the ISPD 2005 benchmark suite [21]. HPWL for the placements generated for the PEKO-MS circuits are proven to be less than 3% above optimal for most cases and within 8% of optimal on all cases.

cs.NI↗

arXiv2023

FPGA-Based In-Vivo Calcium Image Decoding for Closed-Loop Feedback Applications

Miniaturized calcium imaging is an emerging neural recording technique that has been widely used for monitoring neural activity on a large scale at a specific brain region of rats or mice. Most existing calcium-image analysis pipelines operate offline. This results in long processing latency, making it difficult to realize closed-loop feedback stimulation for brain research. In recent work, we have proposed an FPGA-based real-time calcium image processing pipeline for closed-loop feedback applications. It can perform real-time calcium image motion correction, enhancement, fast trace extraction, and real-time decoding from extracted traces. Here, we extend this work by proposing a variety of neural network based methods for real-time decoding and evaluate the tradeoff among these decoding methods and accelerator designs. We introduced the implementation of the neural network based decoders on the FPGA, and showed their speedup against the implementation on the ARM processor. Our FPGA implementation enables the real-time calcium image decoding with sub-ms processing latency for closed-loop feedback applications.

eess.IV↗

arXiv2023

CHARM: Composing Heterogeneous Accelerators for Matrix Multiply on Versal ACAP Architecture

Dense matrix multiply (MM) serves as one of the most heavily used kernels in deep learning applications. To cope with the high computation demands of these applications, heterogeneous architectures featuring both FPGA and dedicated ASIC accelerators have emerged as promising platforms. For example, the AMD/Xilinx Versal ACAP architecture combines general-purpose CPU cores and programmable logic with AI Engine processors optimized for AI/ML. With 400 AIEs, it provides up to 6.4 TFLOPs performance for 32-bit floating-point data. However, machine learning models often contain both large and small MM operations. While large MM operations can be parallelized efficiently across many cores, small MM operations typically cannot. We observe that executing some small MM layers from the BERT natural language processing model on a large, monolithic MM accelerator in Versal ACAP achieved less than 5% of the theoretical peak performance. Therefore, one key question arises: How can we design accelerators to fully use the abundant computation resources under limited communication bandwidth for applications with multiple MM layers of diverse sizes? We identify the biggest system throughput bottleneck resulting from the mismatch of massive computation resources of one monolithic accelerator and the various MM layers of small sizes in the application. To resolve this problem, we propose the CHARM framework to compose multiple diverse MM accelerator architectures working concurrently towards different layers in one application. We deploy the CHARM framework for four different applications, including BERT, ViT, NCF, MLP, on the AMD Versal ACAP VCK190 evaluation board. Our experiments show that we achieve 1.46 TFLOPs, 1.61 TFLOPs, 1.74 TFLOPs, and 2.94 TFLOPs inference throughput for BERT, ViT, NCF and MLP, which obtain 5.40x, 32.51x, 1.00x and 1.00x throughput gains compared to one monolithic accelerator.

cs.AR↗

arXiv2022

Callipepla: Stream Centric Instruction Set and Mixed Precision for Accelerating Conjugate Gradient Solver

The continued growth in the processing power of FPGAs coupled with high bandwidth memories (HBM), makes systems like the Xilinx U280 credible platforms for linear solvers which often dominate the run time of scientific and engineering applications. In this paper, we present Callipepla, an accelerator for a preconditioned conjugate gradient linear solver (CG). FPGA acceleration of CG faces three challenges: (1) how to support an arbitrary problem and terminate acceleration processing on the fly, (2) how to coordinate long-vector data flow among processing modules, and (3) how to save off-chip memory bandwidth and maintain double (FP64) precision accuracy. To tackle the three challenges, we present (1) a stream-centric instruction set for efficient streaming processing and control, (2) vector streaming reuse (VSR) and decentralized vector flow scheduling to coordinate vector data flow among modules and further reduce off-chip memory accesses with a double memory channel design, and (3) a mixed precision scheme to save bandwidth yet still achieve effective double precision quality solutions. To the best of our knowledge, this is the first work to introduce the concept of VSR for data reusing between on-chip modules to reduce unnecessary off-chip accesses for FPGA accelerators. We prototype the accelerator on a Xilinx U280 HBM FPGA. Our evaluation shows that compared to the Xilinx HPC product, the XcgSolver, Callipepla achieves a speedup of 3.94x, 3.36x higher throughput, and 2.94x better energy efficiency. Compared to an NVIDIA A100 GPU which has 4x the memory bandwidth of Callipepla, we still achieve 77% of its throughput with 3.34x higher energy efficiency. The code is available at https://github.com/UCLA-VAST/Callipepla.

cs.AR↗

arXiv2022

Democratizing Domain-Specific Computing

In the past few years, domain-specific accelerators (DSAs), such as Google's Tensor Processing Units, have shown to offer significant performance and energy efficiency over general-purpose CPUs. An important question is whether typical software developers can design and implement their own customized DSAs, with affordability and efficiency, to accelerate their applications. This article presents our answer to this question.

cs.AR↗

arXiv2022

TAPA: A Scalable Task-Parallel Dataflow Programming Framework for Modern FPGAs with Co-Optimization of HLS and Physical Design

In this paper, we propose TAPA, an end-to-end framework that compiles a C++ task-parallel dataflow program into a high-frequency FPGA accelerator. Compared to existing solutions, TAPA has two major advantages. First, TAPA provides a set of convenient APIs that allow users to easily express flexible and complex inter-task communication structures. Second, TAPA adopts a coarse-grained floorplanning step during HLS compilation for accurate pipelining of potential critical paths. In addition, TAPA implements several optimization techniques specifically tailored for modern HBM-based FPGAs. In our experiments with a total of 43 designs, we improve the average frequency from 147 MHz to 297 MHz (a 102% improvement) with no loss of throughput and a negligible change in resource utilization. Notably, in 16 experiments we make the originally unroutable designs achieve 274 MHz on average. The framework is available at https://github.com/UCLA-VAST/tapa and the core floorplan module is available at https://github.com/UCLA-VAST/AutoBridge.

cs.AR↗

arXiv2022

Domain-Specific Quantum Architecture Optimization

With the steady progress in quantum computing over recent years, roadmaps for upscaling quantum processors have relied heavily on the targeted qubit architectures. So far, similarly to the early age of classical computing, these designs have been crafted by human experts. These general-purpose architectures, however, leave room for customization and optimization, especially when targeting popular near-term QC applications. In classical computing, customized architectures have demonstrated significant performance and energy efficiency gains over general-purpose counterparts. In this paper, we present a framework for optimizing quantum architectures, specifically through customizing qubit connectivity. It is the first work that (1) provides performance guarantees by integrating architecture optimization with an optimal compiler, (2) evaluates the impact of connectivity customization under a realistic crosstalk error model, and (3) benchmarks on realistic circuits of near-term interest, such as the quantum approximate optimization algorithm (QAOA) and quantum convolutional neural network (QCNN). We demonstrate up to 59% fidelity improvement in simulation by optimizing the heavy-hexagon architecture for QAOA circuits, and up to 14% improvement on the grid architecture. For the QCNN circuit, architecture optimization improves fidelity by 11% on the heavy-hexagon architecture and 605% on the grid architecture.

cs.AR↗

arXiv2022

TopSort: A High-Performance Two-Phase Sorting Accelerator Optimized on HBM-based FPGAs

The emergence of high-bandwidth memory (HBM) brings new opportunities to boost the performance of sorting acceleration on FPGAs, which was conventionally bounded by the available off-chip memory bandwidth. However, it is nontrivial for designers to fully utilize this immense bandwidth. First, the existing sorter designs cannot be directly scaled at the increasing rate of available off-chip bandwidth, as the required on-chip resource usage grows at a much faster rate and would bound the sorting performance in turn. Second, designers need an in-depth understanding of HBM characteristics to effectively utilize the HBM bandwidth. To tackle these challenges, we present TopSort, a novel two-phase sorting solution optimized for HBM-based FPGAs. In the first phase, 16 merge trees work in parallel to fully utilize 32 HBM channels. In the second phase, TopSort reuses the logic from phase one to form a wider merge tree to merge the partially sorted results from phase one. TopSort also adopts HBM-specific optimizations to reduce resource overhead and improve bandwidth utilization. TopSort can sort up to 4 GB data using all 32 HBM channels, with an overall sorting performance of 15.6 GB/s. TopSort is 6.7x and 2.2x faster than state-of-the-art CPU and FPGA sorters.

cs.AR↗

arXiv2022

Serpens: A High Bandwidth Memory Based Accelerator for General-Purpose Sparse Matrix-Vector Multiplication

Sparse matrix-vector multiplication (SpMV) multiplies a sparse matrix with a dense vector. SpMV plays a crucial role in many applications, from graph analytics to deep learning. The random memory accesses of the sparse matrix make accelerator design challenging. However, high bandwidth memory (HBM) based FPGAs are a good fit for designing accelerators for SpMV. In this paper, we present Serpens, an HBM based accelerator for general-purpose SpMV.Serpens features (1) a general-purpose design, (2) memory-centric processing engines, and (3) index coalescing to support the efficient processing of arbitrary SpMVs. From the evaluation of twelve large-size matrices, Serpens is 1.91x and 1.76x better in terms of geomean throughput than the latest accelerators GraphLiLy and Sextans, respectively. We also evaluate 2,519 SuiteSparse matrices, and Serpens achieves 2.10x higher throughput than a K80 GPU. For the energy/bandwidth efficiency, Serpens is 1.71x/1.99x, 1.90x/2.69x, and 6.25x/4.06x better compared with GraphLily, Sextans, and K80, respectively. After scaling up to 24 HBM channels, Serpens achieves up to 60.55~GFLOP/s (30,204~MTEPS) and up to 3.79x over GraphLily. The code is available at https://github.com/UCLA-VAST/Serpens.

cs.AR↗

arXiv2022

Pyxis: An Open-Source Performance Dataset of Sparse Accelerators

Specialized accelerators provide gains of performance and efficiency in specific domains of applications. Sparse data structures or/and representations exist in a wide range of applications. However, it is challenging to design accelerators for sparse applications because no architecture or performance-level analytic models are able to fully capture the spectrum of the sparse data. Accelerator researchers rely on real execution to get precise feedback for their designs. In this work, we present PYXIS, a performance dataset for specialized accelerators on sparse data. PYXIS collects accelerator designs and real execution performance statistics. Currently, there are 73.8 K instances in PYXIS. PYXIS is open-source, and we are constantly growing PYXIS with new accelerator designs and performance statistics. PYXIS can benefit researchers in the fields of accelerator, architecture, performance, algorithm, and many related topics.

cs.LG↗

arXiv2022

Sextans: A Streaming Accelerator for General-Purpose Sparse-Matrix Dense-Matrix Multiplication

Sparse-Matrix Dense-Matrix multiplication (SpMM) is the key operator for a wide range of applications, including scientific computing, graph processing, and deep learning. Architecting accelerators for SpMM is faced with three challenges - (1) the random memory accessing and unbalanced load in processing because of random distribution of elements in sparse matrices, (2) inefficient data handling of the large matrices which can not be fit on-chip, and (3) anon-general-purpose accelerator design where one accelerator can only process a fixed-size problem. In this paper, we present Sextans, an accelerator for general-purpose SpMM processing. Sextans accelerator features (1) fast random access using on-chip memory, (2) streaming access to off-chip large matrices, (3) PE-aware non-zero scheduling for balanced workload with an II=1 pipeline, and (4) hardware flexibility to enable prototyping the hardware once to support SpMMs of different size as a general-purpose accelerator. We leverage high bandwidth memory (HBM) for the efficient accessing of both sparse and dense matrices. In the evaluation, we present an FPGA prototype Sextans which is executable on a Xilinx U280 HBM FPGA board and a projected prototype Sextans-P with higher bandwidth comparable to V100 and more frequency optimization. We conduct a comprehensive evaluation on 1,400 SpMMs on a wide range of sparse matrices including 50 matrices from SNAP and 150 from SuiteSparse. WecompareSextanswith NVIDIA K80 and V100 GPUs.Sextansachieves a 2.50x geomean speedup over K80 GPU andSextans-Pachieves a 1.14x geomean speedup over V100 GPU (4.94x over K80). The code is available at https://github.com/linghaosong/Sextans.

cs.AR↗

arXiv2021

Search for Optimal Systolic Arrays: A Comprehensive Automated Exploration Framework and Lessons Learned

Systolic arrays have been widely used for accelerating HPC and deep learning applications. There is a plethora of previous works on the performance tuning of systolic arrays, but usually based on a number of oversimplified assumptions (e.g., only considering divisors for loop tiling, pruning based on off-chip data communication) to reduce the design space. In this paper, we present a comprehensive design space exploration tool named Odyssey for systolic array optimization. Odyssey does not rely on artificial assumptions to limit the design space, and yet it is highly efficient and scalable with a hybrid optimization technique. For example, for a 1024x1024x1024 matrix multiplication, it finds designs that reach 90% of the optimal performance in 5 seconds with a single CPU thread. Moreover, using Odyssey, we unveil and quantify the suboptimality introduced by multiple commonly used oversimplifications in prior studies for systolic array design space exploration. For example, Odyssey results show that limiting to divisors for loop tiling leads to a 39% performance loss, and pruning based on off-chip data movement results in a 45% performance loss. We applied Odyssey to explore the architecture trade-offs for matrix multiplication and convolutional neural network, providing inspiration into possible optimizations for these two applications.

cs.AR↗