{"id":8463,"date":"2012-11-06T22:52:42","date_gmt":"2012-11-06T20:52:42","guid":{"rendered":"http:\/\/hgpu.org\/?p=8463"},"modified":"2012-11-06T22:52:42","modified_gmt":"2012-11-06T20:52:42","slug":"adapting-irregular-computations-to-large-cpu-gpu-clusters-in-the-madness-framework","status":"publish","type":"post","link":"https:\/\/hgpu.org\/?p=8463","title":{"rendered":"Adapting Irregular Computations to Large CPU-GPU Clusters in the MADNESS Framework"},"content":{"rendered":"<p>Graphics Processing Units (GPUs) are becoming the workhorse of scalable computations. MADNESS is a scientific framework used especially for computational chemistry. Most MADNESS applications use operators that involve many small tensor computations, resulting in a less regular organization of computations on GPUs. A single GPU kernel may have to multiply by hundreds of small square matrices (with fixed dimension ranging from 10 to 28). We demonstrate a scalable CPU-GPU implementation of the MADNESS framework over a 500-node partition on the Titan supercomputer. For this hybrid CPU-GPU implementation, we observe up to a 2.3-times speedup compared to an equivalent CPU-only implementation with 16 cores per node. For smaller matrices, we demonstrate a speedup of 2.2-times by using a custom CUDA kernel rather than a cuBLAS-based kernel.<\/p>\n","protected":false},"excerpt":{"rendered":"<p>Graphics Processing Units (GPUs) are becoming the workhorse of scalable computations. MADNESS is a scientific framework used especially for computational chemistry. Most MADNESS applications use operators that involve many small tensor computations, resulting in a less regular organization of computations on GPUs. A single GPU kernel may have to multiply by hundreds of small square [&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":[66,11,89,3],"tags":[1790,165,1782,238,14,106,20,1241],"class_list":["post-8463","post","type-post","status-publish","format-standard","hentry","category-chemistry","category-computer-science","category-nvidia-cuda","category-paper","tag-chemistry","tag-computational-chemistry","tag-computer-science","tag-cublas","tag-cuda","tag-gpu-cluster","tag-nvidia","tag-tesla-m2090"],"views":2467,"jetpack_publicize_connections":[],"jetpack_featured_media_url":"","jetpack_sharing_enabled":true,"_links":{"self":[{"href":"https:\/\/hgpu.org\/index.php?rest_route=\/wp\/v2\/posts\/8463","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=8463"}],"version-history":[{"count":0,"href":"https:\/\/hgpu.org\/index.php?rest_route=\/wp\/v2\/posts\/8463\/revisions"}],"wp:attachment":[{"href":"https:\/\/hgpu.org\/index.php?rest_route=%2Fwp%2Fv2%2Fmedia&parent=8463"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/hgpu.org\/index.php?rest_route=%2Fwp%2Fv2%2Fcategories&post=8463"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/hgpu.org\/index.php?rest_route=%2Fwp%2Fv2%2Ftags&post=8463"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}