{"title":"在OpenCL中启用c++ 20 Unseq执行策略","authors":"Po-Yao Chang, Tai-Liang Chen, Jenq-Kuen Lee","doi":"10.1145/3456669.3456674","DOIUrl":null,"url":null,"abstract":"This work facilitates the usage of unsequenced execution policy as seen in C++20 standard library with the newly introduced OpenCL kernel language, C++ for OpenCL. By passing unseq, a global object of type unsequenced_policy, as an argument to selected C++ parallel algorithms, the function would then be vectorized with the help of clang and LLVM. This work complements the introduction of C++ for OpenCL, which brings the core language part of C++17 to OpenCL while leaving out the standard library part. In the best case, we see a whopping 6.9 time speedup.","PeriodicalId":73497,"journal":{"name":"International Workshop on OpenCL","volume":null,"pages":null},"PeriodicalIF":0.0000,"publicationDate":"2021-04-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"2","resultStr":"{\"title\":\"Enabling the Use of C++20 Unseq Execution Policy for OpenCL\",\"authors\":\"Po-Yao Chang, Tai-Liang Chen, Jenq-Kuen Lee\",\"doi\":\"10.1145/3456669.3456674\",\"DOIUrl\":null,\"url\":null,\"abstract\":\"This work facilitates the usage of unsequenced execution policy as seen in C++20 standard library with the newly introduced OpenCL kernel language, C++ for OpenCL. By passing unseq, a global object of type unsequenced_policy, as an argument to selected C++ parallel algorithms, the function would then be vectorized with the help of clang and LLVM. This work complements the introduction of C++ for OpenCL, which brings the core language part of C++17 to OpenCL while leaving out the standard library part. In the best case, we see a whopping 6.9 time speedup.\",\"PeriodicalId\":73497,\"journal\":{\"name\":\"International Workshop on OpenCL\",\"volume\":null,\"pages\":null},\"PeriodicalIF\":0.0000,\"publicationDate\":\"2021-04-27\",\"publicationTypes\":\"Journal Article\",\"fieldsOfStudy\":null,\"isOpenAccess\":false,\"openAccessPdf\":\"\",\"citationCount\":\"2\",\"resultStr\":null,\"platform\":\"Semanticscholar\",\"paperid\":null,\"PeriodicalName\":\"International Workshop on OpenCL\",\"FirstCategoryId\":\"1085\",\"ListUrlMain\":\"https://doi.org/10.1145/3456669.3456674\",\"RegionNum\":0,\"RegionCategory\":null,\"ArticlePicture\":[],\"TitleCN\":null,\"AbstractTextCN\":null,\"PMCID\":null,\"EPubDate\":\"\",\"PubModel\":\"\",\"JCR\":\"\",\"JCRName\":\"\",\"Score\":null,\"Total\":0}","platform":"Semanticscholar","paperid":null,"PeriodicalName":"International Workshop on OpenCL","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/10.1145/3456669.3456674","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"","JCRName":"","Score":null,"Total":0}
引用次数: 2
摘要
这项工作促进了在新引入的OpenCL内核语言c++ for OpenCL中使用c++ 20标准库中的非顺序执行策略。通过传递unseq (unsequenced_policy类型的全局对象)作为所选c++并行算法的参数,该函数将在clang和LLVM的帮助下进行矢量化。这项工作补充了面向OpenCL的c++的介绍,它将c++ 17的核心语言部分引入OpenCL,而忽略了标准库部分。在最好的情况下,我们看到了惊人的6.9的时间加速。
Enabling the Use of C++20 Unseq Execution Policy for OpenCL
This work facilitates the usage of unsequenced execution policy as seen in C++20 standard library with the newly introduced OpenCL kernel language, C++ for OpenCL. By passing unseq, a global object of type unsequenced_policy, as an argument to selected C++ parallel algorithms, the function would then be vectorized with the help of clang and LLVM. This work complements the introduction of C++ for OpenCL, which brings the core language part of C++17 to OpenCL while leaving out the standard library part. In the best case, we see a whopping 6.9 time speedup.