Graphcore

2021 IEEE Hot Chips 33 Symposium (HCS) Pub Date : 2021-08-22 DOI:10.1109/HCS52781.2021.9567075

S. Knowles

{"title":"Graphcore","authors":"S. Knowles","doi":"10.1109/HCS52781.2021.9567075","DOIUrl":null,"url":null,"abstract":"IPU Foundations • AI is nascent – facilitating exploration is as vital as executing known algorithms. • Tera-scale models will be necessary for “super-human” AI. • Sparse evaluation will be necessary at tera-scale, $\\text{for} \\$ $ and Watts. • Rich natural data is sequences, images, and graphs. • Post-Dennard, keep memory and logic close. • Post-Moore, parallel computing over many chips.","PeriodicalId":246531,"journal":{"name":"2021 IEEE Hot Chips 33 Symposium (HCS)","volume":"8 1","pages":"0"},"PeriodicalIF":0.0000,"publicationDate":"2021-08-22","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"43","resultStr":"{\"title\":\"Graphcore\",\"authors\":\"S. Knowles\",\"doi\":\"10.1109/HCS52781.2021.9567075\",\"DOIUrl\":null,\"url\":null,\"abstract\":\"IPU Foundations • AI is nascent – facilitating exploration is as vital as executing known algorithms. • Tera-scale models will be necessary for “super-human” AI. • Sparse evaluation will be necessary at tera-scale, $\\\\text{for} \\\\$ $ and Watts. • Rich natural data is sequences, images, and graphs. • Post-Dennard, keep memory and logic close. • Post-Moore, parallel computing over many chips.\",\"PeriodicalId\":246531,\"journal\":{\"name\":\"2021 IEEE Hot Chips 33 Symposium (HCS)\",\"volume\":\"8 1\",\"pages\":\"0\"},\"PeriodicalIF\":0.0000,\"publicationDate\":\"2021-08-22\",\"publicationTypes\":\"Journal Article\",\"fieldsOfStudy\":null,\"isOpenAccess\":false,\"openAccessPdf\":\"\",\"citationCount\":\"43\",\"resultStr\":null,\"platform\":\"Semanticscholar\",\"paperid\":null,\"PeriodicalName\":\"2021 IEEE Hot Chips 33 Symposium (HCS)\",\"FirstCategoryId\":\"1085\",\"ListUrlMain\":\"https://doi.org/10.1109/HCS52781.2021.9567075\",\"RegionNum\":0,\"RegionCategory\":null,\"ArticlePicture\":[],\"TitleCN\":null,\"AbstractTextCN\":null,\"PMCID\":null,\"EPubDate\":\"\",\"PubModel\":\"\",\"JCR\":\"\",\"JCRName\":\"\",\"Score\":null,\"Total\":0}","platform":"Semanticscholar","paperid":null,"PeriodicalName":"2021 IEEE Hot Chips 33 Symposium (HCS)","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/10.1109/HCS52781.2021.9567075","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"","JCRName":"","Score":null,"Total":0}

引用次数: 43

摘要

•人工智能是新生的——促进探索与执行已知算法一样重要。•“超人类”人工智能将需要万亿级模型。•稀疏评估在兆级，$\text{for} \$ $ $和瓦特级是必要的。•丰富的自然数据是序列，图像和图形。•后登纳德，保持记忆和逻辑接近。•后摩尔时代，多芯片并行计算。

本文章由计算机程序翻译，如有差异，请以英文原文为准。

查看原文本刊更多论文

Graphcore

IPU Foundations • AI is nascent – facilitating exploration is as vital as executing known algorithms. • Tera-scale models will be necessary for “super-human” AI. • Sparse evaluation will be necessary at tera-scale, $\text{for} \$ $ and Watts. • Rich natural data is sequences, images, and graphs. • Post-Dennard, keep memory and logic close. • Post-Moore, parallel computing over many chips.

求助全文

通过发布文献求助，成功后即可免费获取论文全文。去求助

来源期刊

2021 IEEE Hot Chips 33 Symposium (HCS)

自引率

0.00%

发文量