Abstract
The demand for thread-level-parallelism (TLP) on commodity processors is endless as it is essential for gaining performance and saving energy. However, TLP in today's programs is limited by dependences that must be satisfied at run time. We have found that for nondeterministic programs, some of these actual dependences can be satisfied with alternative data that can be generated in parallel, thus boosting the program's TLP. Satisfying these dependences with alternative data nonetheless produces final outputs that match those of the original nondeterministic program. To demonstrate the practicality of our technique, we describe the design, implementation, and evaluation of our compilers, autotuner, profiler, and runtime, which are enabled by our proposed C++ programming language extensions. The resulting system boosts the performance of six well-known nondeterministic and multi-threaded benchmarks by 158.2% (geometric mean) on a 28-core Intel-based platform.
| Original language | English (US) |
|---|---|
| Title of host publication | ASPLOS 2018 - 23rd International Conference on Architectural Support for Programming Languages and Operating Systems |
| Publisher | Association for Computing Machinery |
| Pages | 432-447 |
| Number of pages | 16 |
| ISBN (Electronic) | 9781450349116 |
| DOIs | |
| State | Published - Mar 19 2018 |
| Event | 23rd International Conference on Architectural Support for Programming Languages and Operating Systems, ASPLOS 2018 - Williamsburg, United States Duration: Mar 24 2018 → Mar 28 2018 |
Publication series
| Name | International Conference on Architectural Support for Programming Languages and Operating Systems - ASPLOS |
|---|
Other
| Other | 23rd International Conference on Architectural Support for Programming Languages and Operating Systems, ASPLOS 2018 |
|---|---|
| Country | United States |
| City | Williamsburg |
| Period | 3/24/18 → 3/28/18 |
Fingerprint
Keywords
- Dependences
- Parallelization
- Speculation
ASJC Scopus subject areas
- Software
- Information Systems
- Hardware and Architecture
Cite this
}
Unconventional parallelization of nondeterministic applications. / Deiana, Enrico A.; St Amour, Vincent; Dinda, Peter A; Hardavellas, Nikos; Campanoni, Simone.
ASPLOS 2018 - 23rd International Conference on Architectural Support for Programming Languages and Operating Systems. Association for Computing Machinery, 2018. p. 432-447 (International Conference on Architectural Support for Programming Languages and Operating Systems - ASPLOS).Research output: Chapter in Book/Report/Conference proceeding › Conference contribution
TY - GEN
T1 - Unconventional parallelization of nondeterministic applications
AU - Deiana, Enrico A.
AU - St Amour, Vincent
AU - Dinda, Peter A
AU - Hardavellas, Nikos
AU - Campanoni, Simone
PY - 2018/3/19
Y1 - 2018/3/19
N2 - The demand for thread-level-parallelism (TLP) on commodity processors is endless as it is essential for gaining performance and saving energy. However, TLP in today's programs is limited by dependences that must be satisfied at run time. We have found that for nondeterministic programs, some of these actual dependences can be satisfied with alternative data that can be generated in parallel, thus boosting the program's TLP. Satisfying these dependences with alternative data nonetheless produces final outputs that match those of the original nondeterministic program. To demonstrate the practicality of our technique, we describe the design, implementation, and evaluation of our compilers, autotuner, profiler, and runtime, which are enabled by our proposed C++ programming language extensions. The resulting system boosts the performance of six well-known nondeterministic and multi-threaded benchmarks by 158.2% (geometric mean) on a 28-core Intel-based platform.
AB - The demand for thread-level-parallelism (TLP) on commodity processors is endless as it is essential for gaining performance and saving energy. However, TLP in today's programs is limited by dependences that must be satisfied at run time. We have found that for nondeterministic programs, some of these actual dependences can be satisfied with alternative data that can be generated in parallel, thus boosting the program's TLP. Satisfying these dependences with alternative data nonetheless produces final outputs that match those of the original nondeterministic program. To demonstrate the practicality of our technique, we describe the design, implementation, and evaluation of our compilers, autotuner, profiler, and runtime, which are enabled by our proposed C++ programming language extensions. The resulting system boosts the performance of six well-known nondeterministic and multi-threaded benchmarks by 158.2% (geometric mean) on a 28-core Intel-based platform.
KW - Dependences
KW - Parallelization
KW - Speculation
UR - http://www.scopus.com/inward/record.url?scp=85045390947&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=85045390947&partnerID=8YFLogxK
U2 - 10.1145/3173162.3173181
DO - 10.1145/3173162.3173181
M3 - Conference contribution
AN - SCOPUS:85045390947
T3 - International Conference on Architectural Support for Programming Languages and Operating Systems - ASPLOS
SP - 432
EP - 447
BT - ASPLOS 2018 - 23rd International Conference on Architectural Support for Programming Languages and Operating Systems
PB - Association for Computing Machinery
ER -