ISCA2014
From HELIX
Line 7: | Line 7: | ||
''Proc. International Symposium on Computer Architecture (ISCA), June, 2014'' | ''Proc. International Symposium on Computer Architecture (ISCA), June, 2014'' | ||
+ | <br> | ||
Data dependences in sequential programs limit parallelization because extracted threads cannot run independently. Although thread-level speculation can avoid the need for precise dependence analysis, communication overheads required to synchronize actual dependences counteract the benefits of parallelization. To address these challenges, we propose a lightweight architectural enhancement co-designed with a parallelizing compiler, which together can decouple communication from thread execution. Simulations of these approaches, applied to a processor with 16 Intel Atom-like cores, show an average of 6.85x performance speedup for six SPEC CINT2000 benchmarks. | Data dependences in sequential programs limit parallelization because extracted threads cannot run independently. Although thread-level speculation can avoid the need for precise dependence analysis, communication overheads required to synchronize actual dependences counteract the benefits of parallelization. To address these challenges, we propose a lightweight architectural enhancement co-designed with a parallelizing compiler, which together can decouple communication from thread execution. Simulations of these approaches, applied to a processor with 16 Intel Atom-like cores, show an average of 6.85x performance speedup for six SPEC CINT2000 benchmarks. | ||
[ [[media:ISCA2014_Paper.pdf|Paper]] ] | [ [[media:ISCA2014_Paper.pdf|Paper]] ] |
Revision as of 13:22, 21 April 2014
HELIX-RC: An Architecture-Compiler Co-Design for Automatic Parallelization of Irregular Programs
Simone Campanoni, Kevin Brownell, Svilen Kanev, Timothy Jones, Gu-Yeon Wei, David Brooks
Proc. International Symposium on Computer Architecture (ISCA), June, 2014
Data dependences in sequential programs limit parallelization because extracted threads cannot run independently. Although thread-level speculation can avoid the need for precise dependence analysis, communication overheads required to synchronize actual dependences counteract the benefits of parallelization. To address these challenges, we propose a lightweight architectural enhancement co-designed with a parallelizing compiler, which together can decouple communication from thread execution. Simulations of these approaches, applied to a processor with 16 Intel Atom-like cores, show an average of 6.85x performance speedup for six SPEC CINT2000 benchmarks.
[ Paper ]