{"id":7004,"date":"2012-01-23T01:12:05","date_gmt":"2012-01-22T23:12:05","guid":{"rendered":"http:\/\/hgpu.org\/?p=7004"},"modified":"2012-01-23T01:13:52","modified_gmt":"2012-01-22T23:13:52","slug":"strassens-matrix-multiplication-on-gpus","status":"publish","type":"post","link":"https:\/\/hgpu.org\/?p=7004","title":{"rendered":"Strassen&#8217;s Matrix Multiplication on GPUs"},"content":{"rendered":"<p>We provide efficient single-precision and integer GPU implementations of Strassen&#8217;s algorithm as well as of Winograd&#8217;s variant. On an NVIDIA C1060 GPU, a speedup of 32% (35%) is obtained for Strassen&#8217;s 4-level implementation and 33% (36%) for Winograd&#8217;s variant relative to the sgemm (integer version of sgemm) code in CUBLAS 3.0 when multiplying 16384&#215;16384 matrices. The maximum numerical error for the single-precision implementations is about 2 orders of magnitude higher than those for sgemm when n = 16384 and is zero for the integer implementations.<\/p>\n","protected":false},"excerpt":{"rendered":"<p>We provide efficient single-precision and integer GPU implementations of Strassen&#8217;s algorithm as well as of Winograd&#8217;s variant. On an NVIDIA C1060 GPU, a speedup of 32% (35%) is obtained for Strassen&#8217;s 4-level implementation and 33% (36%) for Winograd&#8217;s variant relative to the sgemm (integer version of sgemm) code in CUBLAS 3.0 when multiplying 16384&#215;16384 matrices. [&hellip;]<\/p>\n","protected":false},"author":351,"featured_media":0,"comment_status":"open","ping_status":"open","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":false,"jetpack_social_options":{"image_generator_settings":{"template":"highway","default_image_id":0,"font":"","enabled":false},"version":2}},"categories":[36,11,89,3],"tags":[1787,1782,14,324,20,199],"class_list":["post-7004","post","type-post","status-publish","format-standard","hentry","category-algorithms","category-computer-science","category-nvidia-cuda","category-paper","tag-algorithms","tag-computer-science","tag-cuda","tag-matrix-multiplication","tag-nvidia","tag-tesla-c1060"],"views":2507,"jetpack_publicize_connections":[],"jetpack_featured_media_url":"","jetpack_sharing_enabled":true,"_links":{"self":[{"href":"https:\/\/hgpu.org\/index.php?rest_route=\/wp\/v2\/posts\/7004","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=7004"}],"version-history":[{"count":1,"href":"https:\/\/hgpu.org\/index.php?rest_route=\/wp\/v2\/posts\/7004\/revisions"}],"predecessor-version":[{"id":7005,"href":"https:\/\/hgpu.org\/index.php?rest_route=\/wp\/v2\/posts\/7004\/revisions\/7005"}],"wp:attachment":[{"href":"https:\/\/hgpu.org\/index.php?rest_route=%2Fwp%2Fv2%2Fmedia&parent=7004"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/hgpu.org\/index.php?rest_route=%2Fwp%2Fv2%2Fcategories&post=7004"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/hgpu.org\/index.php?rest_route=%2Fwp%2Fv2%2Ftags&post=7004"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}