Abstract
In addition to their role as simulation engines, modern supercomputers can be harnessed for scientific visualization. Their extensive concurrency, parallel storage systems, and high-performance interconnects can mitigate the expanding size and complexity of scientific datasets and prepare for in situ visualization of these data. In ongoing research into testing parallel volume rendering on the IBM Blue Gene/P (BG/P), we measure performance of disk I/O, rendering, and compositing on large datasets, and evaluate bottlenecks with respect to system-specific I/O and communication patterns. To extend the scalability of the direct-send image compositing stage of the volume rendering algorithm, we limit the number of compositing cores when many small messages are exchanged. To improve the data-loading stage of the volume renderer, we study the I/O signatures of the algorithm in detail. The results of this research affirm that a distributed-memory computing architecture such as BG/P is a scalable platform for large visualization problems.
| Original language | English (US) |
|---|---|
| Title of host publication | ICPP-2009 - The 38th International Conference on Parallel Processing |
| Pages | 566-573 |
| Number of pages | 8 |
| DOIs | |
| State | Published - Dec 1 2009 |
| Event | 38th International Conference on Parallel Processing, ICPP-2009 - Vienna, Austria Duration: Sep 22 2009 → Sep 25 2009 |
Other
| Other | 38th International Conference on Parallel Processing, ICPP-2009 |
|---|---|
| Country | Austria |
| City | Vienna |
| Period | 9/22/09 → 9/25/09 |
Fingerprint
Keywords
- Distributed scientific visualization
- Image compositing
- Parallel I/O
- Parallel volume rendering
ASJC Scopus subject areas
- Software
- Mathematics(all)
- Hardware and Architecture
Cite this
End-to-end study of parallel volume rendering on the IBM blue Gene/p. / Peterka, Tom; Yu, Hongfeng; Ross, Robert; Ma, Kwan-Liu; Latham, Rob.
ICPP-2009 - The 38th International Conference on Parallel Processing. 2009. p. 566-573 5362481.Research output: Chapter in Book/Report/Conference proceeding › Conference contribution
}
TY - GEN
T1 - End-to-end study of parallel volume rendering on the IBM blue Gene/p
AU - Peterka, Tom
AU - Yu, Hongfeng
AU - Ross, Robert
AU - Ma, Kwan-Liu
AU - Latham, Rob
PY - 2009/12/1
Y1 - 2009/12/1
N2 - In addition to their role as simulation engines, modern supercomputers can be harnessed for scientific visualization. Their extensive concurrency, parallel storage systems, and high-performance interconnects can mitigate the expanding size and complexity of scientific datasets and prepare for in situ visualization of these data. In ongoing research into testing parallel volume rendering on the IBM Blue Gene/P (BG/P), we measure performance of disk I/O, rendering, and compositing on large datasets, and evaluate bottlenecks with respect to system-specific I/O and communication patterns. To extend the scalability of the direct-send image compositing stage of the volume rendering algorithm, we limit the number of compositing cores when many small messages are exchanged. To improve the data-loading stage of the volume renderer, we study the I/O signatures of the algorithm in detail. The results of this research affirm that a distributed-memory computing architecture such as BG/P is a scalable platform for large visualization problems.
AB - In addition to their role as simulation engines, modern supercomputers can be harnessed for scientific visualization. Their extensive concurrency, parallel storage systems, and high-performance interconnects can mitigate the expanding size and complexity of scientific datasets and prepare for in situ visualization of these data. In ongoing research into testing parallel volume rendering on the IBM Blue Gene/P (BG/P), we measure performance of disk I/O, rendering, and compositing on large datasets, and evaluate bottlenecks with respect to system-specific I/O and communication patterns. To extend the scalability of the direct-send image compositing stage of the volume rendering algorithm, we limit the number of compositing cores when many small messages are exchanged. To improve the data-loading stage of the volume renderer, we study the I/O signatures of the algorithm in detail. The results of this research affirm that a distributed-memory computing architecture such as BG/P is a scalable platform for large visualization problems.
KW - Distributed scientific visualization
KW - Image compositing
KW - Parallel I/O
KW - Parallel volume rendering
UR - http://www.scopus.com/inward/record.url?scp=77951496395&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=77951496395&partnerID=8YFLogxK
U2 - 10.1109/ICPP.2009.27
DO - 10.1109/ICPP.2009.27
M3 - Conference contribution
SN - 9780769538020
SP - 566
EP - 573
BT - ICPP-2009 - The 38th International Conference on Parallel Processing
ER -