TY - JOUR UR - http://lib.ugent.be/catalog/pug01:8585865 ID - pug01:8585865 LA - eng TI - RPPM : rapid performance prediction of multithreaded applications on multicore hardware PY - 2018 JO - (2018) IEEE COMPUTER ARCHITECTURE LETTERS SN - 1556-6056 SN - 1556-6064 PB - 2018 AU - De Pestel, Sander UGent 000110586363 802001591947 AU - Van den Steen, Sam UGent 000080508885 802001582954 974744902520 0000-0003-3630-2214 AU - Akram, Shoaib UGent 000111081568 802001224963 AU - Eeckhout, Lieven TW06 801001255603 0000-0001-8792-4473 AB - This paper proposes RPPM which, based on a microarchitecture-independent profile of a multithreaded application, predicts its performance on a previously unseen multicore platform. RPPM breaks up multithreaded program execution into epochs based on synchronization primitives, and then predicts per-epoch active execution times for each thread and synchronization overhead to arrive at a prediction for overall application performance. RPPM predicts performance within 12 percent on average (27 percent max error) compared to cycle-level simulation. We present a case study to illustrate that RPPM can be used for making accurate multicore design trade-offs early in the design cycle. ER -Download RIS file
| 00000nam^a2200301^i^4500 | |||
| 001 | 8585865 | ||
| 005 | 20201002075423.0 | ||
| 008 | 181217s2018------------------------eng-- | ||
| 022 | a 1556-6056 | ||
| 022 | a 1556-6064 | ||
| 024 | a 000440862600021 2 wos | ||
| 024 | a 1854/LU-8585865 2 handle | ||
| 024 | a 10.1109/lca.2018.2849983 2 doi | ||
| 040 | a UGent | ||
| 245 | a RPPM : rapid performance prediction of multithreaded applications on multicore hardware | ||
| 260 | c 2018 | ||
| 520 | a This paper proposes RPPM which, based on a microarchitecture-independent profile of a multithreaded application, predicts its performance on a previously unseen multicore platform. RPPM breaks up multithreaded program execution into epochs based on synchronization primitives, and then predicts per-epoch active execution times for each thread and synchronization overhead to arrive at a prediction for overall application performance. RPPM predicts performance within 12 percent on average (27 percent max error) compared to cycle-level simulation. We present a case study to illustrate that RPPM can be used for making accurate multicore design trade-offs early in the design cycle. | ||
| 598 | a A1 | ||
| 700 | a De Pestel, Sander u UGent 0 000110586363 0 802001591947 0 975693857661 9 37BC6B74-F0EE-11E1-A9DE-61C894A0A6B4 | ||
| 700 | a Van den Steen, Sam u UGent 0 000080508885 0 802001582954 0 974744902520 0 0000-0003-3630-2214 9 13A6AB6E-F0EE-11E1-A9DE-61C894A0A6B4 | ||
| 700 | a Akram, Shoaib u UGent 0 000111081568 0 802001224963 0 972830444438 9 C37B81AE-F0EE-11E1-A197-91C894A0A6B4 | ||
| 700 | a Eeckhout, Lieven u TW06 0 801001255603 0 0000-0001-8792-4473 9 F57D866C-F0ED-11E1-A9DE-61C894A0A6B4 | ||
| 650 | a Technology and Engineering | ||
| 653 | a hardware and architecture | ||
| 653 | a modeling | ||
| 653 | a micro-architecture | ||
| 653 | a performance | ||
| 653 | a multi-threaded | ||
| 773 | t IEEE COMPUTER ARCHITECTURE LETTERS g IEEE Comput. Archit. Lett. 2018. 17 (2) p.183-186 q 17:2<183 | ||
| 856 | 3 fullText u https://biblio.ugent.be/publication/8585865/file/8585866 z [UGent only] y cal2018-RPPM.pdf | ||
| 856 | 3 fullText u https://biblio.ugent.be/publication/8585865/file/8676327 z [open] y 8585865_accepted.pdf | ||
| 920 | a article | ||
| Z30 | x EA 1 TW06 | ||
| 922 | a UGENT-EA | ||
| 856 | 3 fullText u http://sfxit.ugent.be/sfx_local?sid=bellow&atitle=RPPM%20%3A%20rapid%20performance%20prediction%20of%20multithreaded%20applications%20on%20multicore%20hardware&issn=1556-6064&volume=17&issue=2&spage=183&date=2018&svc.fulltext=yes z [ugent] y SFX openurl | ||
All data below are available with an Open Data Commons Open Database License. You are free to copy, distribute and use the database; to produce works from the database; to modify, transform and build upon the database. As long as you attribute the data sets to the source, publish your adapted database with ODbL license, and keep the dataset open (don't use technical measures such as DRM to restrict access to the database).
The datasets are also available as weekly exports.
