Matches in SemOpenAlex for { <https://semopenalex.org/work/W2125979435> ?p ?o ?g. }
Showing items 1 to 91 of
91
with 100 items per page.
- W2125979435 abstract "Current graphics processing units (GPUs) utilize the single instruction multiple thread (SIMT) execution model. With SIMT, a group of logical threads executes such that all threads in the group execute a single common instruction on a particular cycle. To enable control flow to diverge within the group of threads, GPUs partially serialize execution and follow a single control flow path at a time. The execution of the threads in the group that are not on the current path is masked. Most current GPUs rely on a hardware reconvergence stack to track the multiple concurrent paths and to choose a single path for execution. Control flow paths are pushed onto the stack when they diverge and are popped off of the stack to enable threads to reconverge and keep lane utilization high. The stack algorithm guarantees optimal reconvergence for applications with structured control flow as it traverses the structured control-flow tree depth first. The downside of using the reconvergence stack is that only a single path is followed, which does not maximize available parallelism, degrading performance in some cases. We propose a change to the stack hardware in which the execution of two different paths can be interleaved. While this is a fundamental change to the stack concept, we show how dual-path execution can be implemented with only modest changes to current hardware and that parallelism is increased without sacrificing optimal (structured) control-flow reconvergence. We perform a detailed evaluation of a set of benchmarks with divergent control flow and demonstrate that the dual-path stack architecture is much more robust compared to previous approaches for increasing path parallelism. Dual-path execution either matches the performance of the baseline single-path stack architecture or outperforms single-path execution by 14.9% on average and by over 30% in some cases." @default.
- W2125979435 created "2016-06-24" @default.
- W2125979435 creator A5013680653 @default.
- W2125979435 creator A5091648103 @default.
- W2125979435 date "2013-02-01" @default.
- W2125979435 modified "2023-10-18" @default.
- W2125979435 title "The dual-path execution model for efficient GPU control flow" @default.
- W2125979435 cites W1979527452 @default.
- W2125979435 cites W2040469876 @default.
- W2125979435 cites W2047060659 @default.
- W2125979435 cites W2078983643 @default.
- W2125979435 cites W2080592089 @default.
- W2125979435 cites W2090584832 @default.
- W2125979435 cites W2135947393 @default.
- W2125979435 cites W2145866640 @default.
- W2125979435 cites W2148443481 @default.
- W2125979435 cites W2155568054 @default.
- W2125979435 cites W2156540297 @default.
- W2125979435 cites W2156831150 @default.
- W2125979435 cites W2160428323 @default.
- W2125979435 cites W2167675119 @default.
- W2125979435 cites W3148394109 @default.
- W2125979435 doi "https://doi.org/10.1109/hpca.2013.6522352" @default.
- W2125979435 hasPublicationYear "2013" @default.
- W2125979435 type Work @default.
- W2125979435 sameAs 2125979435 @default.
- W2125979435 citedByCount "39" @default.
- W2125979435 countsByYear W21259794352013 @default.
- W2125979435 countsByYear W21259794352014 @default.
- W2125979435 countsByYear W21259794352015 @default.
- W2125979435 countsByYear W21259794352016 @default.
- W2125979435 countsByYear W21259794352017 @default.
- W2125979435 countsByYear W21259794352018 @default.
- W2125979435 countsByYear W21259794352019 @default.
- W2125979435 countsByYear W21259794352020 @default.
- W2125979435 countsByYear W21259794352021 @default.
- W2125979435 countsByYear W21259794352023 @default.
- W2125979435 crossrefType "proceedings-article" @default.
- W2125979435 hasAuthorship W2125979435A5013680653 @default.
- W2125979435 hasAuthorship W2125979435A5091648103 @default.
- W2125979435 hasBestOaLocation W21259794352 @default.
- W2125979435 hasConcept C111919701 @default.
- W2125979435 hasConcept C115874739 @default.
- W2125979435 hasConcept C119024030 @default.
- W2125979435 hasConcept C120314980 @default.
- W2125979435 hasConcept C127413603 @default.
- W2125979435 hasConcept C138101251 @default.
- W2125979435 hasConcept C160191386 @default.
- W2125979435 hasConcept C173608175 @default.
- W2125979435 hasConcept C193702766 @default.
- W2125979435 hasConcept C199360897 @default.
- W2125979435 hasConcept C201995342 @default.
- W2125979435 hasConcept C202491316 @default.
- W2125979435 hasConcept C2777735758 @default.
- W2125979435 hasConcept C41008148 @default.
- W2125979435 hasConcept C52723943 @default.
- W2125979435 hasConcept C9395851 @default.
- W2125979435 hasConceptScore W2125979435C111919701 @default.
- W2125979435 hasConceptScore W2125979435C115874739 @default.
- W2125979435 hasConceptScore W2125979435C119024030 @default.
- W2125979435 hasConceptScore W2125979435C120314980 @default.
- W2125979435 hasConceptScore W2125979435C127413603 @default.
- W2125979435 hasConceptScore W2125979435C138101251 @default.
- W2125979435 hasConceptScore W2125979435C160191386 @default.
- W2125979435 hasConceptScore W2125979435C173608175 @default.
- W2125979435 hasConceptScore W2125979435C193702766 @default.
- W2125979435 hasConceptScore W2125979435C199360897 @default.
- W2125979435 hasConceptScore W2125979435C201995342 @default.
- W2125979435 hasConceptScore W2125979435C202491316 @default.
- W2125979435 hasConceptScore W2125979435C2777735758 @default.
- W2125979435 hasConceptScore W2125979435C41008148 @default.
- W2125979435 hasConceptScore W2125979435C52723943 @default.
- W2125979435 hasConceptScore W2125979435C9395851 @default.
- W2125979435 hasLocation W21259794351 @default.
- W2125979435 hasLocation W21259794352 @default.
- W2125979435 hasOpenAccess W2125979435 @default.
- W2125979435 hasPrimaryLocation W21259794351 @default.
- W2125979435 hasRelatedWork W1563877120 @default.
- W2125979435 hasRelatedWork W2023505575 @default.
- W2125979435 hasRelatedWork W2135947393 @default.
- W2125979435 hasRelatedWork W2155568054 @default.
- W2125979435 hasRelatedWork W2168921806 @default.
- W2125979435 hasRelatedWork W2533181480 @default.
- W2125979435 hasRelatedWork W3037718968 @default.
- W2125979435 hasRelatedWork W3148394109 @default.
- W2125979435 hasRelatedWork W3180108069 @default.
- W2125979435 hasRelatedWork W4293865043 @default.
- W2125979435 isParatext "false" @default.
- W2125979435 isRetracted "false" @default.
- W2125979435 magId "2125979435" @default.
- W2125979435 workType "article" @default.