BEGIN:VCALENDAR
PRODID:-//Microsoft Corporation//Outlook MIMEDIR//EN
VERSION:1.0
BEGIN:VEVENT
DTSTART:20121113T223000Z
DTEND:20121113T230000Z
LOCATION:255-BC
DESCRIPTION;ENCODING=QUOTED-PRINTABLE:ABSTRACT: The growing scale of concurrency requires automated abstraction techniques=0Ato cut down the effort in concurrent system analysis. In this paper, we show that the =0Ahigh degree of behavioral symmetry present in GPU programs=0Aallows CUDA race detection to be dramatically simplified through=0Aabstraction. Our abstraction techniques is one of automatically=0Acreating parametric flows---control-flow equivalence classes of threads that =0Adiverge in the same manner---and checking for data races only across a pair of threads per parametric flow. We have implemented this approach as an extension of our recently=0Aproposed GKLEE symbolic analysis framework and show that all our=0Aprevious results are dramatically improved in that (i) the parametric flow-based=0Aanalysis takes far less time, and (ii) because of the much higher scalability=0Aof the analysis, we can detect even more data race situations that were=0Apreviously missed by GKLEE because it was forced to downscale examples=0Ato limit analysis complexity.
SUMMARY:Parametric Flows - Automated Behavior Equivalencing for Symbolic Analysis of Races in CUDA Programs
PRIORITY:3
END:VEVENT
END:VCALENDAR