اشترك بالحزمة الذهبية واحصل على وصول غير محدود شمرا أكاديميا
تسجيل مستخدم جديدZIPPER: Exploiting Tile- and Operator-level Parallelism for General and Scalable Graph Neural Network Acceleration
212
0
0.0
(
0
)
اسأل ChatGPT حول البحث
ﻻ يوجد ملخص باللغة العربية
Graph neural networks (GNNs) start to gain momentum after showing significant performance improvement in a variety of domains including molecular science, recommendation, and transportation. Turning such performance improvement of GNNs into practical applications relies on effective and efficient execution, especially for inference. However, neither CPU nor GPU can meet these needs if considering both performance and energy efficiency. Thats because accelerating GNNs is challenging due to their excessive memory usage and arbitrary interleaving of diverse operations. Besides, the semantics gap between the high-level GNN programming model and efficient hardware makes it difficult in accelerating general-domain GNNs. To address the challenge, we propose Zipper, an efficient yet general acceleration system for GNNs. The keys to Zipper include a graph-native intermediate representation (IR) and the associated compiler. By capturing GNN primitive operations and representing with GNN IR, Zipper is able to fit GNN semantics into hardware structure for efficient execution. The IR also enables GNN-specific optimizations including sparse graph tiling and redundant operation elimination. We further present an hardware architecture design consisting of dedicated blocks for different primitive operations, along with a run-time scheduler to map a IR program to the hardware blocks. Our evaluation shows that Zipper achieves 93.6x speedup and 147x energy reduction over Intel Xeon CPU, and 1.56x speedup and 4.85x energy reduction over NVIDIA V100 GPU on averages.
قيم البحث
اقرأ أيضاً
This paper summarizes the idea of Subarray-Level Parallelism (SALP) in DRAM, which was published in ISCA 2012, and examines the works significance and future potential. Modern DRAMs have multiple banks to serve multiple memory requests in parallel. H
owever, when two requests go to the same bank, they have to be served serially, exacerbating the high latency of on-chip memory. Adding more banks to the system to mitigate this problem incurs high system cost. Our goal in this work is to achieve the benefits of increasing the number of banks with a low-cost approach. To this end, we propose three new mechanisms, SALP-1, SALP-2, and MASA (Multitude of Activated Subarrays), to reduce the serialization of different requests that go to the same bank. The key observation exploited by our mechanisms is that a modern DRAM bank is implemented as a collection of subarrays that operate largely independently while sharing few global peripheral structures. Our three proposed mechanisms mitigate the negative impact of bank serialization by overlapping different components of the bank access latencies of multiple requests that go to different subarrays within the same bank. SALP-1 requires no changes to the existing DRAM structure, and needs to only reinterpret some of the existing DRAM timing parameters. SALP-2 and MASA require only modest changes (< 0.15% area overhead) to the DRAM peripheral structures, which are much less design constrained than the DRAM core. Our evaluations show that SALP-1, SALP-2 and MASA significantly improve performance for both single-core systems (7%/13%/17%) and multi-core systems (15%/16%/20%), averaged across a wide range of workloads. We also demonstrate that our mechanisms can be combined with application-aware memory request scheduling in multicore systems to further improve performance and fairness.
Exploiting sparsity is a key technique in accelerating quantized convolutional neural network (CNN) inference on mobile devices. Prior sparse CNN accelerators largely exploit un-structured sparsity and achieve significant speedups. Due to the unbound
ed, largely unpredictable sparsity patterns, however, exploiting unstructured sparsity requires complicated hardware design with significant energy and area overhead, which is particularly detrimental to mobile/IoT inference scenarios where energy and area efficiency are crucial. We propose to exploit structured sparsity, more specifically, Density Bound Block (DBB) sparsity for both weights and activations. DBB block tensors bound the maximum number of non-zeros per block. DBB thus exposes statically predictable sparsity patterns that enable lean sparsity-exploiting hardware. We propose new hardware primitives to implement DBB sparsity for (static) weights and (dynamic) activations, respectively, with very low overheads. Building on top of the primitives, we describe S2TA, a systolic array-based CNN accelerator that exploits joint weight and activation DBB sparsity and new dimensions of data reuse unavailable on the traditional systolic array. S2TA in 16nm achieves more than 2x speedup and energy reduction compared to a strong baseline of a systolic array with zero-value clock gating, over five popular CNN benchmarks. Compared to two recent non-systolic sparse accelerators, Eyeriss v2 (65nm) and SparTen (45nm), S2TA in 65nm uses about 2.2x and 3.1x less energy per inference, respectively.
The combination of Winograds algorithm and systolic array architecture has demonstrated the capability of improving DSP efficiency in accelerating convolutional neural networks (CNNs) on FPGA platforms. However, handling arbitrary convolution kernel
sizes in FPGA-based Winograd processing elements and supporting efficient data access remain underexplored. In this work, we are the first to propose an optimized Winograd processing element (WinoPE), which can naturally support multiple convolution kernel sizes with the same amount of computing resources and maintains high runtime DSP efficiency. Using the proposed WinoPE, we construct a highly efficient systolic array accelerator, termed WinoCNN. We also propose a dedicated memory subsystem to optimize the data access. Based on the accelerator architecture, we build accurate resource and performance modeling to explore optimal accelerator configurations under different resource constraints. We implement our proposed accelerator on multiple FPGAs, which outperforms the state-of-the-art designs in terms of both throughput and DSP efficiency. Our implementation achieves DSP efficiency up to 1.33 GOPS/DSP and throughput up to 3.1 TOPS with the Xilinx ZCU102 FPGA. These are 29.1% and 20.0% better than the best solutions reported previously, respectively.
Graph Neural Networks (GNNs) are a new and increasingly popular family of deep neural network architectures to perform learning on graphs. Training them efficiently is challenging due to the irregular nature of graph data. The problem becomes even mo
re challenging when scaling to large graphs that exceed the capacity of single devices. Standard approaches to distributed DNN training, such as data and model parallelism, do not directly apply to GNNs. Instead, two different approaches have emerged in the literature: whole-graph and sample-based training. In this paper, we review and compare the two approaches. Scalability is challenging with both approaches, but we make a case that research should focus on sample-based training since it is a more promising approach. Finally, we review recent systems supporting sample-based training.
Recently, researches have explored the graph neural network (GNN) techniques on text classification, since GNN does well in handling complex structures and preserving global information. However, previous methods based on GNN are mainly faced with th
e practical problems of fixed corpus level graph structure which do not support online testing and high memory consumption. To tackle the problems, we propose a new GNN based model that builds graphs for each input text with global parameters sharing instead of a single graph for the whole corpus. This method removes the burden of dependence between an individual text and entire corpus which support online testing, but still preserve global information. Besides, we build graphs by much smaller windows in the text, which not only extract more local features but also significantly reduce the edge numbers as well as memory consumption. Experiments show that our model outperforms existing models on several text classification datasets even with consuming less memory.
سجل دخول لتتمكن من نشر تعليقات
التعليقات
جاري جلب التعليقات
سجل دخول لتتمكن من متابعة معايير البحث التي قمت باختيارها
