In cloud gaming, servers perform remote rendering on behalf of thin clients. Such a server must deliver sufficient frame rate (at least 30fps) to each of its clients. At the same time, each client desires an immersive experience, and therefore the server should also provide the best graphics quality possible to each client. Statically provisioning time slices of the server GPU for each client suffers from severe underutilization because clients can come and go, and scenes that the clients need rendered can vary greatly in terms of GPU resource usage over time.<\/p>\n
In this work, we present dJay, a utility-maximizing cloud gaming server that dynamically tunes client GPU rendering workloads in order to 1) ensure all clients get satisfactory frame rate, and 2) provide the best possible graphics quality across clients. To accomplish this, we develop three main components. First, we build an online profiler that collects key cost and benefit data, and distills the data into a reusable regression model. Second, we build an online utility optimizer that uses the regression model to tune GPU workloads for better graphics quality. The optimizer solves the Multiple Choice Knapsack problem. We demonstrate dJay on two high quality commercial games, Doom 3 and Fable 3. Our results show that when compared to a static configuration, we can respond much better to peaks and troughs, achieving up to four times the multi-tenant density on a single server while offering clients the best possible graphics quality.<\/p>\n","protected":false},"excerpt":{"rendered":"
In cloud gaming, servers perform remote rendering on behalf of thin clients. Such a server must deliver sufficient frame rate (at least 30fps) to each of its clients. At the same time, each client desires an immersive experience, and therefore the server should also provide the best graphics quality possible to each client. Statically provisioning […]<\/p>\n","protected":false},"featured_media":0,"template":"","meta":{"msr-url-field":"","msr-podcast-episode":"","msrModifiedDate":"","msrModifiedDateEnabled":false,"ep_exclude_from_search":false,"_classifai_error":"","msr-author-ordering":[{"type":"user_nicename","value":"alecw"},{"type":"user_nicename","value":"davidchu"}],"msr_publishername":"ACM - Association for Computing Machinery","msr_publisher_other":"","msr_booktitle":"","msr_chapter":"","msr_edition":"Symposium on Cloud Computing (SoCC)","msr_editors":"","msr_how_published":"","msr_isbn":"","msr_issue":"","msr_journal":"","msr_number":"","msr_organization":"","msr_pages_string":"","msr_page_range_start":"","msr_page_range_end":"","msr_series":"","msr_volume":"","msr_copyright":"\u00a9 ACM. This is the author's version of the work. It is posted here by permission of ACM for your personal use. Not for redistribution. The definitive version can be found at http:\/\/dl.acm.org.","msr_conference_name":"Symposium on Cloud Computing (SoCC)","msr_doi":"10.1145\/2806777.2806942","msr_arxiv_id":"","msr_s2_paper_id":"","msr_mag_id":"","msr_pubmed_id":"","msr_other_authors":"","msr_other_contributors":"","msr_speaker":"","msr_award":"","msr_affiliation":"","msr_institution":"","msr_host":"","msr_version":"","msr_duration":"","msr_original_fields_of_study":"","msr_release_tracker_id":"","msr_s2_match_type":"","msr_citation_count_updated":"","msr_published_date":"2015-08-01","msr_highlight_text":"","msr_notes":"","msr_longbiography":"","msr_publicationurl":"","msr_external_url":"","msr_secondary_video_url":"","msr_conference_url":"","msr_journal_url":"","msr_s2_pdf_url":"","msr_year":0,"msr_citation_count":0,"msr_influential_citations":0,"msr_reference_count":0,"msr_s2_match_confidence":0,"msr_microsoftintellectualproperty":true,"msr_s2_open_access":false,"msr_s2_author_ids":[],"msr_pub_ids":[],"msr_hide_image_in_river":0,"footnotes":""},"msr-research-highlight":[],"research-area":[13563],"msr-publication-type":[193716],"msr-publisher":[],"msr-focus-area":[],"msr-locale":[268875],"msr-post-option":[],"msr-field-of-study":[],"msr-conference":[],"msr-journal":[],"msr-impact-theme":[],"msr-pillar":[],"class_list":["post-242138","msr-research-item","type-msr-research-item","status-publish","hentry","msr-research-area-data-platform-analytics","msr-locale-en_us"],"msr_publishername":"ACM - Association for Computing Machinery","msr_edition":"Symposium on Cloud Computing (SoCC)","msr_affiliation":"","msr_published_date":"2015-08-01","msr_host":"","msr_duration":"","msr_version":"","msr_speaker":"","msr_other_contributors":"","msr_booktitle":"","msr_pages_string":"","msr_chapter":"","msr_isbn":"","msr_journal":"","msr_volume":"","msr_number":"","msr_editors":"","msr_series":"","msr_issue":"","msr_organization":"","msr_how_published":"","msr_notes":"","msr_highlight_text":"","msr_release_tracker_id":"","msr_original_fields_of_study":"","msr_download_urls":"","msr_external_url":"","msr_secondary_video_url":"","msr_longbiography":"","msr_microsoftintellectualproperty":1,"msr_main_download":"242144","msr_publicationurl":"","msr_doi":"10.1145\/2806777.2806942","msr_publication_uploader":[{"type":"file","title":"djay_camera_ready_v2_acmdl","viewUrl":"https:\/\/www.microsoft.com\/en-us\/research\/wp-content\/uploads\/2016\/06\/djay_camera_ready_v2_acmdl.pdf","id":242144,"label_id":0},{"type":"doi","title":"10.1145\/2806777.2806942","viewUrl":false,"id":false,"label_id":0}],"msr_related_uploader":"","msr_citation_count":0,"msr_citation_count_updated":"","msr_s2_paper_id":"","msr_influential_citations":0,"msr_reference_count":0,"msr_arxiv_id":"","msr_s2_author_ids":[],"msr_s2_open_access":false,"msr_s2_pdf_url":null,"msr_attachments":[],"msr-author-ordering":[{"type":"user_nicename","value":"alecw","user_id":30925,"rest_url":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/microsoft-research\/v1\/researchers?person=alecw"},{"type":"user_nicename","value":"davidchu","user_id":31560,"rest_url":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/microsoft-research\/v1\/researchers?person=davidchu"}],"msr_impact_theme":[],"msr_research_lab":[],"msr_event":[],"msr_group":[144899],"msr_project":[236277],"publication":[],"video":[],"msr-tool":[],"msr_publication_type":"inproceedings","related_content":{"projects":[{"ID":236277,"post_title":"Cloud and Edge Mobile Mixed Reality","post_name":"cloud-powered-vr","post_type":"msr-project","post_date":"2016-06-03 13:39:06","post_modified":"2018-07-16 17:32:51","post_status":"publish","permalink":"https:\/\/www.microsoft.com\/en-us\/research\/project\/cloud-powered-vr\/","post_excerpt":"Our goal is to enable mobile devices to deliver fully immersive and untethered mixed reality experiences. We investigate how to enable this experiences in battery-powered devices through leveraging resources available in either nearby edge devices or the cloud. Some of the challenges we study include latency hiding and bandwidth reduction on cloud streaming, pre-rendering and caching virtual worlds, HMD display power efficiency, render and ML offload as well as edge node scheduling of VR\/AR\/MR workloads.…","_links":{"self":[{"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-project\/236277"}]}}]},"_links":{"self":[{"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-research-item\/242138","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-research-item"}],"about":[{"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/types\/msr-research-item"}],"version-history":[{"count":2,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-research-item\/242138\/revisions"}],"predecessor-version":[{"id":523460,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-research-item\/242138\/revisions\/523460"}],"wp:attachment":[{"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/media?parent=242138"}],"wp:term":[{"taxonomy":"msr-research-highlight","embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-research-highlight?post=242138"},{"taxonomy":"msr-research-area","embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/research-area?post=242138"},{"taxonomy":"msr-publication-type","embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-publication-type?post=242138"},{"taxonomy":"msr-publisher","embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-publisher?post=242138"},{"taxonomy":"msr-focus-area","embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-focus-area?post=242138"},{"taxonomy":"msr-locale","embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-locale?post=242138"},{"taxonomy":"msr-post-option","embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-post-option?post=242138"},{"taxonomy":"msr-field-of-study","embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-field-of-study?post=242138"},{"taxonomy":"msr-conference","embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-conference?post=242138"},{"taxonomy":"msr-journal","embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-journal?post=242138"},{"taxonomy":"msr-impact-theme","embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-impact-theme?post=242138"},{"taxonomy":"msr-pillar","embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-pillar?post=242138"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}