ADHA:用于异构架构的自动数据布局框架

2014 23rd International Conference on Parallel Architecture and Compilation (PACT) Pub Date : 2014-07-17 DOI:10.1145/2628071.2628122

Deepak Majeti, Kuldeep S. Meel, R. Barik, Vivek Sarkar

{"title":"ADHA:用于异构架构的自动数据布局框架","authors":"Deepak Majeti, Kuldeep S. Meel, R. Barik, Vivek Sarkar","doi":"10.1145/2628071.2628122","DOIUrl":null,"url":null,"abstract":"Data layouts play a crucial role in determining the performance of a given application running on a given architecture. Existing parallel programming frameworks for both multicore and heterogeneous systems leave the onus of selecting a data layout to the programmer. Therefore, shifting the burden of data layout selection to optimizing compilers can greatly enhance programmer productivity and application performance. In this work, we introduce ADHA: a two-level hierarchal formulation of the data layout problem for modern heterogeneous architectures. We have created a reference implementation of ADHA in the Heterogeneous Habanero-C (H2C) parallel programming system. ADHA shows significant performance benefits of up to 6.92× compared to manually specified layouts for two benchmark programs running on a CPU+GPU heterogeneous platform.","PeriodicalId":263670,"journal":{"name":"2014 23rd International Conference on Parallel Architecture and Compilation (PACT)","volume":"7 1","pages":"0"},"PeriodicalIF":0.0000,"publicationDate":"2014-07-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"3","resultStr":"{\"title\":\"ADHA: Automatic data layout framework for heterogeneous architectures\",\"authors\":\"Deepak Majeti, Kuldeep S. Meel, R. Barik, Vivek Sarkar\",\"doi\":\"10.1145/2628071.2628122\",\"DOIUrl\":null,\"url\":null,\"abstract\":\"Data layouts play a crucial role in determining the performance of a given application running on a given architecture. Existing parallel programming frameworks for both multicore and heterogeneous systems leave the onus of selecting a data layout to the programmer. Therefore, shifting the burden of data layout selection to optimizing compilers can greatly enhance programmer productivity and application performance. In this work, we introduce ADHA: a two-level hierarchal formulation of the data layout problem for modern heterogeneous architectures. We have created a reference implementation of ADHA in the Heterogeneous Habanero-C (H2C) parallel programming system. ADHA shows significant performance benefits of up to 6.92× compared to manually specified layouts for two benchmark programs running on a CPU+GPU heterogeneous platform.\",\"PeriodicalId\":263670,\"journal\":{\"name\":\"2014 23rd International Conference on Parallel Architecture and Compilation (PACT)\",\"volume\":\"7 1\",\"pages\":\"0\"},\"PeriodicalIF\":0.0000,\"publicationDate\":\"2014-07-17\",\"publicationTypes\":\"Journal Article\",\"fieldsOfStudy\":null,\"isOpenAccess\":false,\"openAccessPdf\":\"\",\"citationCount\":\"3\",\"resultStr\":null,\"platform\":\"Semanticscholar\",\"paperid\":null,\"PeriodicalName\":\"2014 23rd International Conference on Parallel Architecture and Compilation (PACT)\",\"FirstCategoryId\":\"1085\",\"ListUrlMain\":\"https://doi.org/10.1145/2628071.2628122\",\"RegionNum\":0,\"RegionCategory\":null,\"ArticlePicture\":[],\"TitleCN\":null,\"AbstractTextCN\":null,\"PMCID\":null,\"EPubDate\":\"\",\"PubModel\":\"\",\"JCR\":\"\",\"JCRName\":\"\",\"Score\":null,\"Total\":0}","platform":"Semanticscholar","paperid":null,"PeriodicalName":"2014 23rd International Conference on Parallel Architecture and Compilation (PACT)","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/10.1145/2628071.2628122","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"","JCRName":"","Score":null,"Total":0}

引用次数: 3

摘要

数据布局在确定在给定体系结构上运行的给定应用程序的性能方面起着至关重要的作用。现有的多核和异构系统并行编程框架都把选择数据布局的责任留给了程序员。因此，将数据布局选择的负担转移到优化编译器上可以极大地提高程序员的生产力和应用程序的性能。在这项工作中，我们介绍了ADHA:现代异构架构的数据布局问题的两级层次结构。我们在异构Habanero-C (H2C)并行编程系统中创建了ADHA的参考实现。对于运行在CPU+GPU异构平台上的两个基准测试程序，与手动指定的布局相比，ADHA显示出高达6.92倍的显著性能优势。

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

ADHA: Automatic data layout framework for heterogeneous architectures

Data layouts play a crucial role in determining the performance of a given application running on a given architecture. Existing parallel programming frameworks for both multicore and heterogeneous systems leave the onus of selecting a data layout to the programmer. Therefore, shifting the burden of data layout selection to optimizing compilers can greatly enhance programmer productivity and application performance. In this work, we introduce ADHA: a two-level hierarchal formulation of the data layout problem for modern heterogeneous architectures. We have created a reference implementation of ADHA in the Heterogeneous Habanero-C (H2C) parallel programming system. ADHA shows significant performance benefits of up to 6.92× compared to manually specified layouts for two benchmark programs running on a CPU+GPU heterogeneous platform.

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

2014 23rd International Conference on Parallel Architecture and Compilation (PACT)

自引率

0.00%

发文量