skip to main content
US FlagAn official website of the United States government
dot gov icon
Official websites use .gov
A .gov website belongs to an official government organization in the United States.
https lock icon
Secure .gov websites use HTTPS
A lock ( lock ) or https:// means you've safely connected to the .gov website. Share sensitive information only on official, secure websites.

Explore Research Products in the PAR
It may take a few hours for recently added research products to appear in PAR search results.


Title: A Data Prefetcher-Based 1000-Core RISC-V Processor for Efficient Processing of Graph Neural Networks
Graphs-based neural networks have seen tremendous adoption to perform complex predictive analytics on massive real-world graphs. The trend in hardware acceleration has identified significant challenges with harnessing graph locality and workload imbalance due to ultrasparse and irregular matrix computations at a massively parallel scale. State-of-the-art hardware accelerators utilize massive multithreading and asynchronous execution in GPUs to achieve parallel performance at high power consumption. This paper aims to bridge the power-performance gap using the energy efficiency-centric RISC-V ecosystem. A 1000-core RISC-V processor is proposed to unlock massive parallelism in the graphs-based matrix operators to achieve a low-latency data access paradigm in hardware to achieve robust power-performance scaling. Each core implements a single-threaded pipeline with a novel graph-aware data prefetcher at the 1000 cores scale to deliver an average 20× performance per watt advantage over state-of-the-art NVIDIA GPU.  more » « less
Award ID(s):
2429516
PAR ID:
10662388
Author(s) / Creator(s):
 
Publisher / Repository:
IEEE
Date Published:
Journal Name:
IEEE Computer Architecture Letters
Volume:
24
Issue:
1
ISSN:
1556-6056
Page Range / eLocation ID:
73 to 76
Format(s):
Medium: X
Sponsoring Org:
National Science Foundation
More Like this
  1. (, Eurographics Symposium on Parallel Graphics and Visualization)
    Bujack, Roxana and (Ed.)
    Large scale graphs are used to encode data from a variety of application domains such as social networks, the web, biological networks, road maps, and finance. Computing enriching layouts and interactive rendering play an important role in many of these applications. However, producing an efficient and interactive visualization of large graphs remains a major challenge, particularly in the web-browser. Existing state of the art web-based visualization systems such as D3.js, Stardust, and NetV.js struggle to achieve interactive layout and visualization for large scale graphs. In this work, we leverage the latest WebGPU technology to develop GraphWaGu, the first WebGPU-based graph visualization system. WebGPU is a new graphics API that brings the full capabilities of modern GPUs to the web browser. Leveraging the computational capabilities of the GPU using this technology enables GraphWaGu to scale to larger graphs than existing technologies. GraphWaGu embodies both fast parallel rendering and layout creation using modified Frutcherman-Reingold and Barnes-Hut algorithms implemented in WebGPU compute shaders. Experimental results demonstrate that our solution achieves the best performance, scalability, and layout quality when compared to current state of the art web-based graph visualization libraries. All of our source code for the project is available at https://github.com/harp-lab/GraphWaGu. 
    more » « less
  2. ; ; (, International Conference on Computer-Aided Design)
    null (Ed.)
    Recent spectral graph sparsification techniques have shown promising performance in accelerating many numerical and graph algorithms, such as iterative methods for solving large sparse matrices, spectral partitioning of undirected graphs, vectorless verification of power/thermal grids, representation learning of large graphs, etc. However, prior spectral graph sparsification methods rely on fast Laplacian matrix solvers that are usually challenging to implement in practice. This work, for the first time, introduces a solver-free approach (SF-GRASS) for spectral graph sparsification by leveraging emerging spectral graph coarsening and graph signal processing (GSP) techniques. We introduce a local spectral embedding scheme for efficiently identifying spectrally-critical edges that are key to preserving graph spectral properties, such as the first few Laplacian eigenvalues and eigenvectors. Since the key kernel functions in SF-GRASS can be efficiently implemented using sparse-matrix-vector-multiplications (SpMVs), the proposed spectral approach is simple to implement and inherently parallel friendly. Our extensive experimental results show that the proposed method can produce a hierarchy of high-quality spectral sparsifiers in nearly-linear time for a variety of real-world, large-scale graphs and circuit networks when compared with prior state-of-the-art spectral methods. 
    more » « less
  3. ; ; ; ; (, 2019 20th International Workshop on Microprocessor/SoC Test, Security and Verification (MTV))
    Attacks which combine software vulnerabilities and hardware vulnerabilities are emerging security problems. Although the runtime verification or remote attestation can determine the correctness of a system, existing methods suffer from inflexible security policy setup and high performance overheads. Meanwhile, they rarely focus on addressing the threat in the RISC-V architecture, which provides an open Instruction Set Architecture (ISA) of the processsor. In this paper, we propose a comprehensive software and hardware co-verification method to protect the entire RISC-V system in the runtime. The proposed method adopts the Dynamic Information Flow Tracking (DIFT) framework to implement a new Verifier and Prover security architecture for supporting runtime software and hardware coverification. We realize a FPGA prototype on the Rocket-Chip, an RISC-V open-source processor core. The framework is implemented as a co-processor which do not change the architecture of main processor core and the new security architecture can be integrated with other RISC-V processors. 
    more » « less
  4. (, Arizona State University)
    This dissertation introduces a series of digital CIM circuits and architectures that significantly improve power, performance, and area (PPA) metrics for data-intensive workloads. It begins with a programmable CIM design that balances the flexibility of Central-Processing-Units(CPUs)/Graphics Processing Units(GPUs) with the efficiency of ASICs, enabling a broad class of applications. A prototype 28nm CMOS chip is then presented to accelerate general matrix-matrix multiplications (GEMMs) across various fixed-point precisions. The focus then shifts to sparse GEMM acceleration. The first design demonistrates how CIM tailored for channel decoders leverages both fixed and unstructured sparsity to outperform conventional designs. The second design, fabricated in 28nm CMOS, supports diverse unstructured sparse formats and integer precisions, efficiently targeting highly sparse deep neural networks (DNNs). The final design achieves state-of-the-art efficiency in compressed sparse GEMMs, supporting both integer and floating-point data types using shared hardware. It also integrates a RISC-V CPU to manage computation across diverse matrix sizes and model types. Together, these contributions advance CIM as a scalable and efficient platform for future AI and data-centric systems. 
    more » « less
  5. ; ; ; ; (, 2021 IEEE 28th International Conference on High Performance Computing, Data, and Analytics (HiPC))
    Alternating Least Square (ALS) is a classic algorithm to solve matrix factorization widely used in recommendation systems. Existing efforts focus on parallelizing ALS on multi-/many-core platforms to handle large datasets. Recently, an optimized ALS variant called eALS was proposed, and it yields significantly lower time complexity and higher recommending accuracy than ALS. However, it is challenging to parallelize eALS on modern parallel architectures (e.g., CPUs and GPUs) because: 1) eALS’ data dependence prevents it from fine-grained parallel execution, thus eALS cannot fully utilize GPU's massive parallelism, 2) the sparsity of input data causes poor data locality and unbalanced workload, and 3) its large memory usage cannot fit into GPU's limited on-device memory, particularly for real-world large datasets. This paper proposes an efficient CPU/GPU heterogeneous recommendation system based on fast eALS for the first time (called HEALS) that consists of a set of system optimizations. HEALS employs newly designed architecture-adaptive data formats to achieve load balance and good data locality on CPU and GPU. HEALS also presents a CPU/GPU collaboration model that can explore both task parallelism and data parallelism. HEALS also optimizes this collaboration model with data communication overlapping and dynamic workload partition between CPU and GPU. Moreover, HEALS is further enhanced by various parallel techniques (e.g., loop unrolling, vectorization, and GPU parallel reduction). Evaluation results show that HEALS outperforms other state-of-the-art baselines in both performance and recommendation quality. Particularly, HEALS achieves up to 5.75 x better performance than a state-of-the-art ALS GPU library. This work also demonstrates the possibility of conducting fast recommendations on large datasets with constrained (or relaxed) hardware resources, e.g, a single CPU/GPU node. 
    more » « less
  • Citation Formats
  • MLA
  • APA
  • Chicago
  • BibTeX
  • Text Encoded Conversion
  • XML
  • Save / Share this Record
  • Send to Email