PowerPoint Presentation - Efficient Dynamic Verification ：PowerPoint演示文稿的高效动态验证.pptVIP

下载本文档

1
0
约2.5万字
约 30页
2020-08-18 发布于湖北
举报
版权申诉

PowerPoint Presentation - Efficient Dynamic Verification ：PowerPoint演示文稿的高效动态验证.ppt

1、原创力文档（book118）网站文档一经付费（服务费），不意味着购买了该文档的版权，仅供个人/单位学习、研究之用，不得用于商业用途，未经授权，严禁复制、发行、汇编、翻译或者网络传播等，侵权必究。。
2、本站所有内容均由合作方或网友上传，本站不对文档的完整性、权威性及其观点立场正确性做任何保证或承诺！文档内容仅供研究参考，付费前请自行鉴别。如您付费，意味着您自己接受本站规则且自行承担风险，本站不退款、不进行额外附加服务；查看《如何避免下载的几个坑》。如果您已付费下载过本站文档，您可以点击这里二次下载。
3、如文档侵犯商业秘密、侵犯著作权、侵犯人身权等，请点击“版权申诉”（推荐），也可以打举报电话：400-050-0827(电话支持时间：9:00-18:30)。
4、该文档为VIP文档，如果想要下载，成为VIP会员后，下载免费。
5、成为VIP后，下载本文档将扣除1次下载权益。下载后，不支持退款、换文档。如有疑问请联系我们。
6、成为VIP后，您将拥有八大权益，权益包括：VIP文档下载权益、阅读免打扰、文档格式转换、高级专利检索、专属身份标志、高级客服、多端互通、版权登记。
7、VIP文档为合作方或网友上传，每下载1次，网站将根据用户上传文档的质量评分、类型等，对文档贡献者给予高额补贴、流量扶持。如果你也想贡献VIP文档。上传文档

* * * no processor can do all kinds of computation efficiently. For a variety of reasons, acquisition cost should not exceed $200M and the annual operating power cost should not exceed $20M, which limits the power to 20 MW Power wall, cpu is not efficient for all the kinds of computation. No one device can do every computation in an efficient way DARPA study identify four challenge for exascale computing : 1) energy and power challenge 2) memory and storage challenge 3) concurrency and locality challenge 4) resiliency challenge number one issue is energy will not be a practical path to exascale computing a heterogeneous mix of distinct scalar and vector units. o the efforts of NVIDIA and their industry partners. Intels ongoing plans for its Many Integrated Core (MIC) co-processor and AMDs introduction of its CPU-GPU Fusion processors are yet additional indications that the industry is moving to an architecture where trade-off between flexibility / programmability / genericity and efficiency as important as ever today, with transformation of Moore’s Law from speed-based to concurrency-based, due to power considerations One interpretation of “Amdalh’s law” ? We will always need powerful, general purpose cores to speed up sequential parts of our applications! “Future processors will be a mix of general purpose and specialized cores” For this paper, I will mainly talk about the programmability and efficiency, and mainly about how to schedule SPMD computation on heterhgonesou cluster * There is also the scalablity issue but didn’t have much discuss in this presentatoin. ; multi-core, coprocessor, accelerator, are heterogeneous architecture Accelerator are attached to gpu with pci-e bus Programmability is another issue * These programming models very different from each other because they are designed for different devices, To make the parallel program for these heterogenesou devices you can Either use a hybrid programming model, or use a uniform programming