CRII: SHF: Design and Analysis of Processing-Near-Memory Enabled GPU Architecture

CRII：SHF：支持近内存处理的 GPU 架构的设计和分析

• 批准号: 1657336
• 负责人: Adwait Jog
• 金额: $ 17.5万
• 依托单位: College of William and Mary
• 依托单位国家: 美国
• 项目类别: Standard Grant
• 财政年份: 2017
• 资助国家: 美国
• 起止时间: 2017-02-01 至 2020-01-31
• 项目状态: 已结题
• 来源: https://www.nsf.gov/awardsearch/showAward?AWD_ID=1657336&HistoricalAwards=false
• 关键词: CRII; SHF; Design; Analysis; Processing

Graphics Processing Units (GPUs) are becoming an inevitable part of every computing system because of their ability to enable orders of magnitude faster and energy-efficient execution. However, the necessary and continuous scaling of GPUs in terms of performance and energy efficiency will not be an easy task. Prior works have shown that two biggest impediments towards this scaling are the limited memory bandwidth and the excessive data movement across different levels of the memory hierarchy. In order to alleviate these two issues, die-stacking technology is gaining momentum in the realm of high-performance energy-efficient GPU computing. This technology not only enables very high memory bandwidth for better performance but also provides support for processing-near-memory (PNM) to reduce data movement, access latencies, and energy consumption. Although these technologies seem promising, the architectural support and execution models for PNM-based GPUs and their implications on the entire system design have largely been unexplored. This project takes a fresh look at the design and execution model of a PNM-enabled GPU, which consists of multiple memory stacks and each memory stack incorporates a 3D-stacked logic layer that can consist of multiple PNM GPU cores and other uncore components. Considering that GPUs are becoming an inevitable part of every computing system ranging from warehouse-scale computers to wearable devices, the insights resulting from this research can have a long-term positive impact on the GPU-based computing. The findings of this research will be incorporated to existing and new undergraduate and graduate courses, which will directly help in educating and training students, including women and students from diverse backgrounds and minority groups.First, a detailed design space exploration will be performed, which will involve the study of the impact and interactions of different design choices related to PNM cores (e.g., register file, SIMD width, pipeline components, warp occupancy), uncore components at the logic layer (e.g., caches) and stacked memory (e.g., number of stacked memories). Second, a computation distribution framework (CDF) will be developed that will answer: a) when is it preferable to map computations to PNM cores, b) which PNM cores and computations they should be?, and c) how can we effectively take advantage of both PNM and regular GPU cores? The CDF will leverage different static and runtime strategies to address many of such similar questions to push the envelopes of energy efficiency and performance even further. The proposed research components will be evaluated via a wide-range of GPGPU applications.  If successful, the findings of this research would better equip PNM-enabled GPUs to effectively alleviate the two major bottlenecks: memory bandwidth and energy.

图形处理单元 (GPU) 正在成为每个计算系统不可避免的一部分，因为它们能够实现速度提高和节能的数量级。但是，GPU 在性能和能源效率方面的必要和持续扩展不会得到满足。先前的工作表明，这种扩展的两个最大障碍是有限的内存带宽和跨不同级别的内存层次结构的过多数据移动，为了缓解这两个问题，芯片堆叠技术正在取得进展。该技术不仅能够实现极高的内存带宽以实现更好的性能，而且还提供对近内存处理 (PNM) 的支持，以减少数据移动、访问延迟和能耗。尽管这些技术看起来很有前途，但基于 PNM 的 GPU 的架构支持和执行模型及其对整个系统设计的影响在很大程度上尚未得到探索。该项目重新审视了 GPU 的设计和执行模型。支持 PNM 的 GPU，由多个内存堆栈组成，每个内存堆栈都包含一个 3D 堆栈逻辑层，该逻辑层可以由多个 PNM GPU 核心和其他非核心组件组成 考虑到 GPU 是每个计算系统中不可避免的一部分，从仓库到将计算机扩展到可穿戴设备，这项研究得出的见解可以对基于 GPU 的计算产生长期的积极影响。这项研究的结果将被纳入现有和新的本科生和研究生课程中。直接帮助教育和培训学生，包括来自不同背景和少数群体的女性和学生。首先，将进行详细的设计空间探索，其中将涉及与 PNM 核心相关的不同设计选择（例如、寄存器文件、SIMD 宽度、流水线组件、扭曲占用）、逻辑层的非核心组件（例如缓存）和堆栈内存（例如堆栈内存的数量）。将开发一个计算分布框架（CDF），它将回答：a）什么时候最好将计算映射到 PNM 核心，b）它们应该是哪些 PNM 核心和计算？以及 c）我们如何有效地利用两者PNM 和常规 GPU 核心？ CDF 将利用不同的静态和运行时策略来解决许多此类类似问题，以进一步突破能源效率和性能的极限。如果成功的话，这项研究的结果将更好地装备支持 PNM 的 GPU，以有效缓解内存带宽和能耗这两个主要瓶颈。

项目成果

Address-stride assisted approximate load value prediction in GPUs

GPU 中的地址跨距辅助近似负载值预测

• DOI: 10.1145/3330345.3330362
• 发表时间: 2019-06
• 期刊: ICS '19: Proceedings of the ACM International Conference on Supercomputing
• 作者: Wang, Haonan;Ibrahim, Mohamed;Mittal, Sparsh;Jog, Adwait
• 通讯作者: Jog, Adwait

Design and Analysis of Soft-Error Resilience Mechanisms for GPU Register File

GPU寄存器文件软错误恢复机制的设计与分析

• DOI: 10.1109/vlsid.2017.14
• 发表时间: 2024-09-13
• 期刊: 2017 30th International Conference on VLSI Design and 2017 16th International Conference on Embedded Systems (VLSID)
• 作者: Sparsh Mittal;Haonan Wang;Adwait Jog;J. Vetter
• 通讯作者: J. Vetter

Efficient and Fair Multi-programming in GPUs via Effective Bandwidth Management

• DOI: 10.1109/hpca.2018.00030
• 发表时间: 2018-02-01
• 期刊: 2018 IEEE International Symposium on High Performance Computer Architecture (HPCA)
• 作者: Haonan Wang;Fan Luo;M. Ibrahim;Onur Kayiran;Adwait Jog
• 通讯作者: Adwait Jog

Architectural Support for Efficient Large-Scale Automata Processing

高效大规模自动机处理的架构支持

• DOI: 10.1109/micro.2018.00078
• 发表时间: 2018-10
• 期刊: 51st Annual IEEE/ACM International Symposium on Microarchitecture (MICRO
• 作者: Liu, Hongyuan;Ibrahim, Mohamed;Kayiran, Onur;Pai, Sreepathi;Jog, Adwait
• 通讯作者: Jog, Adwait

Opportunistic computing in GPU architectures

GPU 架构中的机会计算

• DOI: 10.1145/3307650.3322212
• 发表时间: 2019-06
• 期刊: ISCA '19: Proceedings of the 46th International Symposium on Computer Architecture
• 作者: Pattnaik, Ashutosh;Tang, Xulong;Kayiran, Onur;Jog, Adwait;Mishra, Asit;Kandemir, Mahmut T.;Sivasubramaniam, Anand;Das, Chita R.
• 通讯作者: Das, Chita R.