Nowadays, Graphics Processing Units (GPUs) dominate in a wide spectrum of computing realms and multi-task is increasingly applied in various complicated applications. To gain higher performance, multi-task programs require cumbersome programming efforts to take advantage of inter-kernel concurrency at source-code level. Although there exist works automatically scheduling kernels to enable inter-kernel concurrency, they all inevitably introduce new programming frameworks and some even bring significant performance downgrade compared to the expertise-based optimizations. To address this issue, we propose KeSCo, a compiler-based scheduler to expose kernel level concurrency in multi-task programs with trivial code modification. In compilation, KeSCo applies a strategy to schedule kernels in task queues, accounting for both load balance and synchronization cost. Also, KeSCo utilizes a customized algorithm designed for computational flow to remove redundant synchronizations. The design is further extended to support multiprocess scenario, where multiple GPU processes are sharing a single context. Evaluations on representative benchmarks show that the proposed approach gains a 1.28x average speedup for multi-task scenario (1.22x for multi-process). Even with lessened programming efforts, our proposed design outperforms two state-of-the-arts GrSched and Taskflow by 1.31x and 1.16x on average, respectively.<\/p>\n","protected":false},"excerpt":{"rendered":"
Nowadays, Graphics Processing Units (GPUs) dominate in a wide spectrum of computing realms and multi-task is increasingly applied in various complicated applications. To gain higher performance, multi-task programs require cumbersome programming efforts to take advantage of inter-kernel concurrency at source-code level. Although there exist works automatically scheduling kernels to enable inter-kernel concurrency, they all inevitably […]<\/p>\n","protected":false},"author":351,"featured_media":0,"comment_status":"open","ping_status":"closed","sticky":false,"template":"","format":"standard","meta":{"_jetpack_memberships_contains_paid_content":false,"footnotes":"","jetpack_publicize_message":"","jetpack_publicize_feature_enabled":true,"jetpack_social_post_already_shared":true,"jetpack_social_options":{"image_generator_settings":{"template":"highway","default_image_id":0,"font":"","enabled":false},"version":2}},"categories":[11,89,3],"tags":[451,955,1782,14,20,2066,854],"class_list":["post-28896","post","type-post","status-publish","format-standard","hentry","category-computer-science","category-nvidia-cuda","category-paper","tag-benchmarking","tag-compilers","tag-computer-science","tag-cuda","tag-nvidia","tag-nvidia-a100","tag-task-scheduling"],"views":1595,"jetpack_publicize_connections":[],"jetpack_featured_media_url":"","jetpack_sharing_enabled":true,"_links":{"self":[{"href":"https:\/\/hgpu.org\/index.php?rest_route=\/wp\/v2\/posts\/28896","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/hgpu.org\/index.php?rest_route=\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/hgpu.org\/index.php?rest_route=\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/hgpu.org\/index.php?rest_route=\/wp\/v2\/users\/351"}],"replies":[{"embeddable":true,"href":"https:\/\/hgpu.org\/index.php?rest_route=%2Fwp%2Fv2%2Fcomments&post=28896"}],"version-history":[{"count":0,"href":"https:\/\/hgpu.org\/index.php?rest_route=\/wp\/v2\/posts\/28896\/revisions"}],"wp:attachment":[{"href":"https:\/\/hgpu.org\/index.php?rest_route=%2Fwp%2Fv2%2Fmedia&parent=28896"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/hgpu.org\/index.php?rest_route=%2Fwp%2Fv2%2Fcategories&post=28896"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/hgpu.org\/index.php?rest_route=%2Fwp%2Fv2%2Ftags&post=28896"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}