{"328071":{"#nid":"328071","#data":{"type":"event","title":"Ph.D. Proposal by Naila Farooqui","body":[{"value":"\u003Cp\u003ETitle: \u003Cstrong\u003EDynamic Instrumentation for Resource Management and Optimization on Heterogeneous CPU\/GPU Platforms\u003C\/strong\u003E\u003C\/p\u003E\u003Cp\u003E\u003Cstrong\u003ENaila Farooqui\u003C\/strong\u003E\u003Cbr \/\u003ESchool of Computer Science\u003Cbr \/\u003ECollege of Computing\u003Cbr \/\u003EGeorgia Institute of Technology\u003C\/p\u003E\u003Cp\u003EDate: October 2nd, 2014 (Thursday)\u003Cbr \/\u003ETime: 12:00 PM - 2:00 PM (ET)\u003Cbr \/\u003ELocation: KACB 1315\u003C\/p\u003E\u003Cp\u003ECommittee:\u003Cbr \/\u003E---------------\u003Cbr \/\u003EDr. Karsten Schwan (Advisor, School of Computer Science, Georgia Tech)\u003Cbr \/\u003EDr. Sudhakar Yalamanchili (School of Electrical and Computer Engineering, Georgia Tech)\u003Cbr \/\u003EDr. Ada Gavrilovska (School of Computer Science, Georgia Tech)\u003Cbr \/\u003EDr. Richard Vuduc (School of Computational Science and Engineering, Georgia Tech)\u003Cbr \/\u003EDr. Vanish Talwar (Systems Research, HP Labs)\u003C\/p\u003E\u003Cp\u003E\u003Cstrong\u003EAbstract:\u003C\/strong\u003E\u003Cbr \/\u003E------------\u003Cbr \/\u003EHeterogeneous parallel architectures like those comprised of CPUs and GPUs\u003Cbr \/\u003Eare a tantalizing compute fabric for performance-hungry developers. While \u003Cbr \/\u003Ethese platforms enable order-of-magnitude performance increases for many \u003Cbr \/\u003Edata-parallel application domains, there remain several open challenges:\u003Cbr \/\u003E(i) the distinct execution models inherent in the heterogeneous devices \u003Cbr \/\u003Epresent on such platforms drives the need to dynamically match workload \u003Cbr \/\u003Echaracteristics to the underlying resources, (ii) the complex architecture \u003Cbr \/\u003Eand programming models of such systems require substantial application\u003Cbr \/\u003Eknowledge and effort-intensive program tuning to achieve high performance, \u003Cbr \/\u003Eand (iii) as such platforms become prevalent, there is a need to extend \u003Cbr \/\u003Etheir utility from running known regular data-parallel applications to \u003Cbr \/\u003Ethe broader set of input-dependent, irregular applications common in \u003Cbr \/\u003Eenterprise settings.\u003Cbr \/\u003EThe key contribution of our research is to employ dynamic instrumentation \u003Cbr \/\u003Eto drive profile-driven resource management and optimizations for such \u003Cbr \/\u003Eheterogeneous hybrid CPU\/GPU platforms, in order to enable high application \u003Cbr \/\u003Eperformance and system throughput. Towards this end, this research will: \u003Cbr \/\u003E(a) enable dynamic instrumentation for GPU-based parallel architectures, \u003Cbr \/\u003Especifically targeting the complex Single-Instruction Multiple-Data (SIMD) \u003Cbr \/\u003Eexecution model, to gain real-time introspection into application behavior; \u003Cbr \/\u003E(b) leverage such dynamic performance data to support novel online resource \u003Cbr \/\u003Emanagement methods that improve application performance and system throughput, \u003Cbr \/\u003Eparticularly for irregular, input-dependent applications; and (c) automate some \u003Cbr \/\u003Eof the programmer effort required to exercise specialized architectural \u003Cbr \/\u003Efeatures of such platforms via instrumentation-driven dynamic code optimizations.\u003C\/p\u003E","summary":null,"format":"limited_html"}],"field_subtitle":"","field_summary":"","field_summary_sentence":[{"value":"Dynamic Instrumentation for Resource Management and Optimization on Heterogeneous CPU\/GPU Platforms"}],"uid":"28077","created_gmt":"2014-09-23 15:34:18","changed_gmt":"2016-10-08 02:09:28","author":"Danielle Ramirez","boilerplate_text":"","field_publication":"","field_article_url":"","field_event_time":{"event_time_start":"2014-10-02T13:00:00-04:00","event_time_end":"2014-10-02T15:00:00-04:00","event_time_end_last":"2014-10-02T15:00:00-04:00","gmt_time_start":"2014-10-02 17:00:00","gmt_time_end":"2014-10-02 19:00:00","gmt_time_end_last":"2014-10-02 19:00:00","rrule":null,"timezone":"America\/New_York"},"extras":[],"groups":[{"id":"221981","name":"Graduate Studies"}],"categories":[],"keywords":[{"id":"2523","name":"cs"},{"id":"1808","name":"graduate students"},{"id":"102851","name":"Phd proposal"}],"core_research_areas":[],"news_room_topics":[],"event_categories":[{"id":"1788","name":"Other\/Miscellaneous"}],"invited_audience":[{"id":"78771","name":"Public"}],"affiliations":[],"classification":[],"areas_of_expertise":[],"news_and_recent_appearances":[],"phone":[],"contact":[],"email":[],"slides":[],"orientation":[],"userdata":""}}}