Abstract
Multicore computers have been widely included in cluster systems. They are shared memory architecture. However, previous research on parallel loop selfscheduling did not consider the feature of multicore computers. It is more suitable for shared-memory multiprocessors to adopt OpenMP for parallel programming. Therefore, in this paper, we propose to adopt hybrid programming model MPI+OpenMP to design loop self-scheduling schemes for cluster systems with multicore computers. Initially, each computer runs only one MPI process no matter how many cores it has. A MPI process will fork OpenMP threads depending on the number of cores in the computer. Each idle slave MPI-process will request tasks from the master process. The tasks dispatched to a process will be executed in parallel by OpenMP threads. According to the experimental results, our method outperforms the previous work by 18.66% or 29.76% depending on the problem size. Moreover, the performance improvement is very stable no matter our method is based on which traditional scheme.
| Original language | English |
|---|---|
| Title of host publication | Proceedings of the 3rd IEEE Asia-Pacific Services Computing Conference, APSCC 2008 |
| Publisher | IEEE Computer Society |
| Pages | 251-256 |
| Number of pages | 6 |
| ISBN (Print) | 9780769534732 |
| DOIs | |
| Publication status | Published - 2008 Jan 1 |
| Event | 3rd IEEE Asia-Pacific Services Computing Conference, APSCC 2008 - Yilan, Taiwan Duration: 2008 Dec 9 → 2008 Dec 12 |
Other
| Other | 3rd IEEE Asia-Pacific Services Computing Conference, APSCC 2008 |
|---|---|
| Country | Taiwan |
| City | Yilan |
| Period | 08-12-09 → 08-12-12 |
Fingerprint
All Science Journal Classification (ASJC) codes
- Computer Science Applications
- Software
- Electrical and Electronic Engineering
Cite this
}
Parallel loop self-scheduling for heterogeneous cluster systems with multi-core computers. / Wu, Chao-Chin; Lai, Lien-Fu; Chiu, Po Hsun.
Proceedings of the 3rd IEEE Asia-Pacific Services Computing Conference, APSCC 2008. IEEE Computer Society, 2008. p. 251-256 4780684.Research output: Chapter in Book/Report/Conference proceeding › Conference contribution
TY - GEN
T1 - Parallel loop self-scheduling for heterogeneous cluster systems with multi-core computers
AU - Wu, Chao-Chin
AU - Lai, Lien-Fu
AU - Chiu, Po Hsun
PY - 2008/1/1
Y1 - 2008/1/1
N2 - Multicore computers have been widely included in cluster systems. They are shared memory architecture. However, previous research on parallel loop selfscheduling did not consider the feature of multicore computers. It is more suitable for shared-memory multiprocessors to adopt OpenMP for parallel programming. Therefore, in this paper, we propose to adopt hybrid programming model MPI+OpenMP to design loop self-scheduling schemes for cluster systems with multicore computers. Initially, each computer runs only one MPI process no matter how many cores it has. A MPI process will fork OpenMP threads depending on the number of cores in the computer. Each idle slave MPI-process will request tasks from the master process. The tasks dispatched to a process will be executed in parallel by OpenMP threads. According to the experimental results, our method outperforms the previous work by 18.66% or 29.76% depending on the problem size. Moreover, the performance improvement is very stable no matter our method is based on which traditional scheme.
AB - Multicore computers have been widely included in cluster systems. They are shared memory architecture. However, previous research on parallel loop selfscheduling did not consider the feature of multicore computers. It is more suitable for shared-memory multiprocessors to adopt OpenMP for parallel programming. Therefore, in this paper, we propose to adopt hybrid programming model MPI+OpenMP to design loop self-scheduling schemes for cluster systems with multicore computers. Initially, each computer runs only one MPI process no matter how many cores it has. A MPI process will fork OpenMP threads depending on the number of cores in the computer. Each idle slave MPI-process will request tasks from the master process. The tasks dispatched to a process will be executed in parallel by OpenMP threads. According to the experimental results, our method outperforms the previous work by 18.66% or 29.76% depending on the problem size. Moreover, the performance improvement is very stable no matter our method is based on which traditional scheme.
UR - http://www.scopus.com/inward/record.url?scp=67049116287&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=67049116287&partnerID=8YFLogxK
U2 - 10.1109/APSCC.2008.166
DO - 10.1109/APSCC.2008.166
M3 - Conference contribution
AN - SCOPUS:67049116287
SN - 9780769534732
SP - 251
EP - 256
BT - Proceedings of the 3rd IEEE Asia-Pacific Services Computing Conference, APSCC 2008
PB - IEEE Computer Society
ER -