Task graph parallelism has emerged as an important tool to efficiently execute large machine learning workloads on GPUs. Users describe a GPU workload in a task dependency graph rather than aggregated GPU operations and dependencies, allowing the runtime to run whole-graph scheduling optimization to significantly improve the performance. While the new CUDA graph execution model has demonstrated significant success on this front, the counterpart for SYCL, a general-purpose heterogeneous programming model using standard C++, remains nascent. Unlike CUDA graph, SYCL runtime leverages out-of-order queues to implicitly create a task execution graph induced by data dependencies. For explicit task dependencies, users are responsible for creating SYCL events and synchronizing them at a non-negligible cost. Furthermore, there is no specialized graph execution model that allows users to offload a task graph directly onto a SYCL device in a similar way to CUDA graph. This paper conducts an experimental study of SYCL’s default task graph parallelism by comparing it with CUDA graph on large-scale machine learning workloads in the recent HPEC Graph Challenge. Our result highlights the need for a new SYCL graph execution model in the standard.<\/p>\n","protected":false},"excerpt":{"rendered":"
Task graph parallelism has emerged as an important tool to efficiently execute large machine learning workloads on GPUs. Users describe a GPU workload in a task dependency graph rather than aggregated GPU operations and dependencies, allowing the runtime to run whole-graph scheduling optimization to significantly improve the performance. While the new CUDA graph execution model […]<\/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":[1782,14,452,1025,20,2088,1845],"class_list":["post-25626","post","type-post","status-publish","format-standard","hentry","category-computer-science","category-nvidia-cuda","category-paper","tag-computer-science","tag-cuda","tag-heterogeneous-systems","tag-machine-learning","tag-nvidia","tag-nvidia-geforce-gtx-2080","tag-sycl"],"views":2164,"jetpack_publicize_connections":[],"jetpack_featured_media_url":"","jetpack_sharing_enabled":true,"_links":{"self":[{"href":"https:\/\/hgpu.org\/index.php?rest_route=\/wp\/v2\/posts\/25626","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=25626"}],"version-history":[{"count":0,"href":"https:\/\/hgpu.org\/index.php?rest_route=\/wp\/v2\/posts\/25626\/revisions"}],"wp:attachment":[{"href":"https:\/\/hgpu.org\/index.php?rest_route=%2Fwp%2Fv2%2Fmedia&parent=25626"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/hgpu.org\/index.php?rest_route=%2Fwp%2Fv2%2Fcategories&post=25626"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/hgpu.org\/index.php?rest_route=%2Fwp%2Fv2%2Ftags&post=25626"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}