3D GPU architecture using cache stacking: Performance, cost, power and thermal analysis

2009 IEEE International Conference on Computer Design Pub Date : 2009-10-04 DOI:10.1109/ICCD.2009.5413147

Ahmed Al-Maashri, Guangyu Sun, Xiangyu Dong, V. Narayanan, Yuan Xie

引用次数: 43

Abstract

Graphics Processing Units (GPUs) offer tremendous computational and processing power. The architecture requires high communication bandwidth and lower latency between computation units and caches. 3D die-stacking technology is a promising approach to meet such requirements. To the best of our knowledge no other study has investigated the implementation of 3D technology in GPUs. In this paper, we study the impact of stacking caches using the 3D technology on GPU performance. We also investigate the benefits of using 3D stacked MRAM on GPUs. Our work includes cost, power, and thermal analysis of the proposed architectural designs. Our results show a 53% geometric mean performance speedup for iso-cycle time architectures and about 19% for iso-cost architectures.

查看原文

微信好友朋友圈 QQ好友复制链接

本刊更多论文

使用缓存堆叠的3D GPU架构:性能、成本、功耗和热分析

图形处理单元(gpu)提供了巨大的计算和处理能力。该架构要求计算单元和缓存之间具有较高的通信带宽和较低的延迟。3D模堆技术是满足这一要求的一种很有前途的方法。据我们所知，没有其他研究调查了gpu中3D技术的实现。在本文中，我们研究了使用3D技术堆叠缓存对GPU性能的影响。我们还研究了在gpu上使用3D堆叠MRAM的好处。我们的工作包括对提议的建筑设计进行成本、功率和热分析。我们的结果表明，等周期时间架构的几何平均性能提升了53%，等成本架构的几何平均性能提升了19%。

本文章由计算机程序翻译，如有差异，请以英文原文为准。

求助全文

约1分钟内获得全文去求助

来源期刊

2009 IEEE International Conference on Computer Design

自引率

0.00%

发文量

期刊最新文献

Empirical performance models for 3T1D memories A novel SoC architecture on FPGA for ultra fast face detection A Technology-Agnostic Simulation Environment (TASE) for iterative custom IC design across processes Low-overhead error detection for Networks-on-Chip Interconnect performance corners considering crosstalk noise