{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,3,6]],"date-time":"2026-03-06T10:41:51Z","timestamp":1772793711391,"version":"3.50.1"},"reference-count":36,"publisher":"Institute of Electrical and Electronics Engineers (IEEE)","issue":"7","license":[{"start":{"date-parts":[[2016,7,1]],"date-time":"2016-07-01T00:00:00Z","timestamp":1467331200000},"content-version":"vor","delay-in-days":0,"URL":"https:\/\/ieeexplore.ieee.org\/Xplorehelp\/downloads\/license-information\/IEEE.html"}],"funder":[{"DOI":"10.13039\/501100001809","name":"National Natural Science Foundation of China","doi-asserted-by":"publisher","award":["61300005"],"award-info":[{"award-number":["61300005"]}],"id":[{"id":"10.13039\/501100001809","id-type":"DOI","asserted-by":"publisher"}]},{"name":"Human-Centered Cyber-Physical Systems Programme at the Advanced Digital Sciences Center from Singapore\u2019s Agency for Science, Technology and Research (A*STAR) and CFAR, one of six centers of STARnet"},{"DOI":"10.13039\/100000028","name":"Semiconductor Research Corporation program sponsored by MARCO and DARPA","doi-asserted-by":"publisher","id":[{"id":"10.13039\/100000028","id-type":"DOI","asserted-by":"publisher"}]}],"content-domain":{"domain":[],"crossmark-restriction":false},"short-container-title":["IEEE Trans. Comput.-Aided Des. Integr. Circuits Syst."],"published-print":{"date-parts":[[2016,7]]},"DOI":"10.1109\/tcad.2015.2501303","type":"journal-article","created":{"date-parts":[[2015,11,17]],"date-time":"2015-11-17T19:36:38Z","timestamp":1447788998000},"page":"1165-1178","source":"Crossref","is-referenced-by-count":17,"title":["An Accurate GPU Performance Model for Effective Control Flow Divergence Optimization"],"prefix":"10.1109","volume":"35","author":[{"given":"Yun","family":"Liang","sequence":"first","affiliation":[]},{"given":"Muhammad Teguh","family":"Satria","sequence":"additional","affiliation":[]},{"ORCID":"https:\/\/orcid.org\/0000-0003-2908-2225","authenticated-orcid":false,"given":"Kyle","family":"Rupnow","sequence":"additional","affiliation":[]},{"given":"Deming","family":"Chen","sequence":"additional","affiliation":[]}],"member":"263","reference":[{"key":"ref33","first-page":"591","article-title":"The dual-path execution model for efficient GPU control flow","author":"rhu","year":"2013","journal-title":"Proc IEEE 19th Int Symp High Perform Comput Archit (HPCA)"},{"key":"ref32","doi-asserted-by":"publisher","DOI":"10.1109\/TPDS.2014.2313342"},{"key":"ref31","doi-asserted-by":"publisher","DOI":"10.1145\/2593069.2593208"},{"key":"ref30","first-page":"461","article-title":"Register and thread structure optimization for GPUs","author":"liang","year":"2013","journal-title":"Proc Asia South Pacific Design Autom Conf (ASP-DAC)"},{"key":"ref36","doi-asserted-by":"publisher","DOI":"10.1109\/MICRO.2014.48"},{"key":"ref35","doi-asserted-by":"publisher","DOI":"10.1145\/2485922.2485953"},{"key":"ref34","doi-asserted-by":"publisher","DOI":"10.1145\/2485922.2485954"},{"key":"ref10","doi-asserted-by":"publisher","DOI":"10.1145\/1555754.1555775"},{"key":"ref11","doi-asserted-by":"publisher","DOI":"10.1109\/ISPASS.2010.5452013"},{"key":"ref12","doi-asserted-by":"publisher","DOI":"10.1145\/1543753.1543756"},{"key":"ref13","year":"0","journal-title":"NVIDIA CUDA Programming guide"},{"key":"ref14","article-title":"Program optimization strategies for data-parallel many-core processors","author":"ryoo","year":"2008"},{"key":"ref15","doi-asserted-by":"publisher","DOI":"10.1145\/37402.37422"},{"key":"ref16","doi-asserted-by":"publisher","DOI":"10.1109\/IPDPS.2009.5161066"},{"key":"ref17","year":"0","journal-title":"CUDA-EC NVIDIA Tesla Bio Bench"},{"key":"ref18","doi-asserted-by":"publisher","DOI":"10.1109\/ICIAFS.2010.5715660"},{"key":"ref19","doi-asserted-by":"publisher","DOI":"10.1186\/1471-2105-8-474"},{"key":"ref28","doi-asserted-by":"publisher","DOI":"10.1109\/HPCA.2015.7056023"},{"key":"ref4","doi-asserted-by":"publisher","DOI":"10.1145\/1693453.1693470"},{"key":"ref27","doi-asserted-by":"publisher","DOI":"10.1109\/ICCAD.2013.6691165"},{"key":"ref3","first-page":"832","article-title":"Real-time implementation and performance optimization of 3D sound localization on GPUs","author":"liang","year":"2012","journal-title":"Proc Design and Test in Europe (DATE) Conf"},{"key":"ref6","doi-asserted-by":"publisher","DOI":"10.1145\/1815961.1815992"},{"key":"ref29","doi-asserted-by":"publisher","DOI":"10.1145\/2830772.2830813"},{"key":"ref5","doi-asserted-by":"publisher","DOI":"10.1109\/MICRO.2007.30"},{"key":"ref8","doi-asserted-by":"publisher","DOI":"10.1145\/1950365.1950408"},{"key":"ref7","doi-asserted-by":"publisher","DOI":"10.1145\/1810085.1810104"},{"key":"ref2","doi-asserted-by":"publisher","DOI":"10.1145\/1687399.1687501"},{"key":"ref9","year":"0","journal-title":"NVIDIA Visual Profiler"},{"key":"ref1","doi-asserted-by":"publisher","DOI":"10.7873\/DATE.2013.279"},{"key":"ref20","doi-asserted-by":"publisher","DOI":"10.1109\/ICCV.2011.6126416"},{"key":"ref22","doi-asserted-by":"publisher","DOI":"10.1145\/2555243.2555285"},{"key":"ref21","doi-asserted-by":"publisher","DOI":"10.1145\/1806596.1806606"},{"key":"ref24","doi-asserted-by":"publisher","DOI":"10.1109\/IPDPS.2012.18"},{"key":"ref23","doi-asserted-by":"publisher","DOI":"10.1109\/HPCSim.2014.6903670"},{"key":"ref26","doi-asserted-by":"publisher","DOI":"10.1145\/2000064.2000093"},{"key":"ref25","doi-asserted-by":"publisher","DOI":"10.1145\/2155620.2155656"}],"container-title":["IEEE Transactions on Computer-Aided Design of Integrated Circuits and Systems"],"original-title":[],"link":[{"URL":"http:\/\/xplorestaging.ieee.org\/ielx7\/43\/7493714\/07329967.pdf?arnumber=7329967","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2021,10,11]],"date-time":"2021-10-11T02:34:57Z","timestamp":1633919697000},"score":1,"resource":{"primary":{"URL":"http:\/\/ieeexplore.ieee.org\/document\/7329967\/"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2016,7]]},"references-count":36,"journal-issue":{"issue":"7"},"URL":"https:\/\/doi.org\/10.1109\/tcad.2015.2501303","relation":{},"ISSN":["0278-0070","1937-4151"],"issn-type":[{"value":"0278-0070","type":"print"},{"value":"1937-4151","type":"electronic"}],"subject":[],"published":{"date-parts":[[2016,7]]}}}