Abstract
Visualizing three-dimensional unstructured data from aerodynamics calculations is challenging because the associated meshes are typically large in size and irregular in both shape and resolution. The goal of this research is to develop a fast, efficient parallel volume rendering algorithm for massively parallel distributed-memory supercomputers consisting of a large number of very powerful processors. We use cell-projection instead of ray-casting to provide maximum flexibility in the data distribution and rendering steps. Effective static load balancing is achieved with a round robin distribution of data cells among the processors. A spatial partitioning tree is used to guide the rendering, optimize the image compositing step, and reduce memory consumption. Communication cost is reduced by buffering messages and by overlapping communication with rendering calculations as much as possible. Tests on the IBM SP2 demonstrate that these strategies provide high rendering rates and good scalability. For a dataset containing half a million tetrahedral cells, we achieve two frames per second for a 400×400-pixel image using 128 processors.
Original language | English (US) |
---|---|
Pages | 95-104 |
Number of pages | 10 |
State | Published - Dec 1 1997 |
Externally published | Yes |
Event | Proceedings of the 1997 IEEE Symposium on Parallel Rendering, PRS - Phoenix, AZ, USA Duration: Oct 20 1997 → Oct 21 1997 |
Other
Other | Proceedings of the 1997 IEEE Symposium on Parallel Rendering, PRS |
---|---|
City | Phoenix, AZ, USA |
Period | 10/20/97 → 10/21/97 |
Fingerprint
ASJC Scopus subject areas
- Computer Science(all)
- Engineering(all)
Cite this
Scalable parallel cell-projection volume rendering algorithm for three-dimensional unstructured data. / Ma, Kwan-Liu; Crockett, Thomas W.
1997. 95-104 Paper presented at Proceedings of the 1997 IEEE Symposium on Parallel Rendering, PRS, Phoenix, AZ, USA, .Research output: Contribution to conference › Paper
}
TY - CONF
T1 - Scalable parallel cell-projection volume rendering algorithm for three-dimensional unstructured data
AU - Ma, Kwan-Liu
AU - Crockett, Thomas W.
PY - 1997/12/1
Y1 - 1997/12/1
N2 - Visualizing three-dimensional unstructured data from aerodynamics calculations is challenging because the associated meshes are typically large in size and irregular in both shape and resolution. The goal of this research is to develop a fast, efficient parallel volume rendering algorithm for massively parallel distributed-memory supercomputers consisting of a large number of very powerful processors. We use cell-projection instead of ray-casting to provide maximum flexibility in the data distribution and rendering steps. Effective static load balancing is achieved with a round robin distribution of data cells among the processors. A spatial partitioning tree is used to guide the rendering, optimize the image compositing step, and reduce memory consumption. Communication cost is reduced by buffering messages and by overlapping communication with rendering calculations as much as possible. Tests on the IBM SP2 demonstrate that these strategies provide high rendering rates and good scalability. For a dataset containing half a million tetrahedral cells, we achieve two frames per second for a 400×400-pixel image using 128 processors.
AB - Visualizing three-dimensional unstructured data from aerodynamics calculations is challenging because the associated meshes are typically large in size and irregular in both shape and resolution. The goal of this research is to develop a fast, efficient parallel volume rendering algorithm for massively parallel distributed-memory supercomputers consisting of a large number of very powerful processors. We use cell-projection instead of ray-casting to provide maximum flexibility in the data distribution and rendering steps. Effective static load balancing is achieved with a round robin distribution of data cells among the processors. A spatial partitioning tree is used to guide the rendering, optimize the image compositing step, and reduce memory consumption. Communication cost is reduced by buffering messages and by overlapping communication with rendering calculations as much as possible. Tests on the IBM SP2 demonstrate that these strategies provide high rendering rates and good scalability. For a dataset containing half a million tetrahedral cells, we achieve two frames per second for a 400×400-pixel image using 128 processors.
UR - http://www.scopus.com/inward/record.url?scp=0031373379&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=0031373379&partnerID=8YFLogxK
M3 - Paper
AN - SCOPUS:0031373379
SP - 95
EP - 104
ER -