{"id":18911,"date":"2019-05-26T13:33:30","date_gmt":"2019-05-26T10:33:30","guid":{"rendered":"https:\/\/hgpu.org\/?p=18911"},"modified":"2019-05-26T13:33:30","modified_gmt":"2019-05-26T10:33:30","slug":"multi-gpu-rendering-with-vulkan-api","status":"publish","type":"post","link":"https:\/\/hgpu.org\/?p=18911","title":{"rendered":"Multi-GPU Rendering with Vulkan API"},"content":{"rendered":"<p>Vulkan API provides a low level interface to modern Graphics Processing Units (GPUs). With this thesis, we demonstrate how to use Vulkan to send commands explicitly to separate GPUs for implementing platform- and vendor independent multi-GPU rendering. We describe how to implement the sort-first and sort-last approaches to perform parallel rendering with Vulkan. We introduce an abstraction library which we have made available, and an application for multi-GPU rendering of meshes. The introduced solution is the first publicly available implementation of heterogeneous multi-GPU rendering with Vulkan API. The introduced abstraction library supports creating partial renderers for the sort-first and sort-last approaches to multi-GPU rendering, and takes care of the details of multi-GPU synchronization and compositing. Performance benchmarks have been performed in order to evaluate the implementation. The sort-last implementation has been tested to render geometry with high-resolution textures, which would otherwise not fit in the memory of a single GPU.<\/p>\n","protected":false},"excerpt":{"rendered":"<p>Vulkan API provides a low level interface to modern Graphics Processing Units (GPUs). With this thesis, we demonstrate how to use Vulkan to send commands explicitly to separate GPUs for implementing platform- and vendor independent multi-GPU rendering. We describe how to implement the sort-first and sort-last approaches to perform parallel rendering with Vulkan. We introduce [&hellip;]<\/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":[11,3],"tags":[2025,7,1782,452,20,1898,1779,176,390,1847],"class_list":["post-18911","post","type-post","status-publish","format-standard","hentry","category-computer-science","category-paper","tag-amd-radeon-rx-460","tag-ati","tag-computer-science","tag-heterogeneous-systems","tag-nvidia","tag-nvidia-geforce-gtx-1080","tag-nvidia-geforce-gtx-970","tag-package","tag-thesis","tag-vulkan"],"views":5476,"jetpack_publicize_connections":[],"jetpack_featured_media_url":"","jetpack_sharing_enabled":true,"_links":{"self":[{"href":"https:\/\/hgpu.org\/index.php?rest_route=\/wp\/v2\/posts\/18911","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=18911"}],"version-history":[{"count":0,"href":"https:\/\/hgpu.org\/index.php?rest_route=\/wp\/v2\/posts\/18911\/revisions"}],"wp:attachment":[{"href":"https:\/\/hgpu.org\/index.php?rest_route=%2Fwp%2Fv2%2Fmedia&parent=18911"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/hgpu.org\/index.php?rest_route=%2Fwp%2Fv2%2Fcategories&post=18911"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/hgpu.org\/index.php?rest_route=%2Fwp%2Fv2%2Ftags&post=18911"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}