On the Correctness of the SIMT Execution Model of GPUs

  • GPUs are becoming a primary resource of computing power. They use a single instruction, multiple threads (SIMT) execution model that executes batches of threads in lockstep. If the control flow of threads within the same batch diverges, the different execution paths are scheduled sequentially; once the control flows reconverge, all threads are executed in lockstep again. Several thread batching mechanisms have been proposed, albeit without establishing their semantic validity or their scheduling properties. To increase the level of confidence in the correctness of GPU-accelerated programs, we formalize the SIMT execution model for a stack-based reconvergence mechanism in an operational semantics and prove its correctness by constructing a simulation between the SIMT semantics and a standard interleaved thread semantics. We also demonstrate that the SIMT execution model produces unfair schedules in some cases. We discuss the problem of unfairness for different batching mechanisms likeGPUs are becoming a primary resource of computing power. They use a single instruction, multiple threads (SIMT) execution model that executes batches of threads in lockstep. If the control flow of threads within the same batch diverges, the different execution paths are scheduled sequentially; once the control flows reconverge, all threads are executed in lockstep again. Several thread batching mechanisms have been proposed, albeit without establishing their semantic validity or their scheduling properties. To increase the level of confidence in the correctness of GPU-accelerated programs, we formalize the SIMT execution model for a stack-based reconvergence mechanism in an operational semantics and prove its correctness by constructing a simulation between the SIMT semantics and a standard interleaved thread semantics. We also demonstrate that the SIMT execution model produces unfair schedules in some cases. We discuss the problem of unfairness for different batching mechanisms like dynamic warp formation and a stack-less reconvergence strategy.show moreshow less

Download full text files

Export metadata

Statistics

Number of document requests

Additional Services

Share in Twitter Search Google Scholar
Metadaten
Author:Axel Habermaier, Alexander KnappORCiDGND
URN:urn:nbn:de:bvb:384-opus4-12719
Frontdoor URLhttps://opus.bibliothek.uni-augsburg.de/opus4/1600
Series (Serial Number):Reports / Technische Berichte der Fakultät für Angewandte Informatik der Universität Augsburg (2012-01)
Type:Report
Language:English
Publishing Institution:Universität Augsburg
Release Date:2012/01/12
Tag:SIMT; Ausführungsmodell
execution model
GND-Keyword:CUDA <Informatik>; Formale Semantik
Institutes:Fakultät für Angewandte Informatik
Fakultät für Angewandte Informatik / Institut für Informatik
Fakultät für Angewandte Informatik / Institut für Software & Systems Engineering
Fakultät für Angewandte Informatik / Institut für Informatik / Lehrstuhl für Softwaretechnik
Fakultät für Angewandte Informatik / Institut für Informatik / Lehrstuhl für Softwaretechnik / Professur für die Grundlagen des Software & Systems Engineering
Dewey Decimal Classification:0 Informatik, Informationswissenschaft, allgemeine Werke / 00 Informatik, Wissen, Systeme / 004 Datenverarbeitung; Informatik
Licence (German):Deutsches Urheberrecht