2000
DOI: 10.1109/5992.814651
|View full text |Cite
|
Sign up to set email alerts
|

Strategies for adopting FVTD on multicomputers [finite-volume time-domain analysis]

Help me understand this report

Search citation statements

Order By: Relevance

Paper Sections

Select...
3
1
1

Citation Types

0
8
0

Year Published

2003
2003
2019
2019

Publication Types

Select...
4
2
1

Relationship

0
7

Authors

Journals

citations
Cited by 10 publications
(8 citation statements)
references
References 12 publications
0
8
0
Order By: Relevance
“…For comparison purposes, we also include the speedups of the same code running on an IBM SP2 machine using MPI in Figure 3. The MPI code uses a three-dimensional decomposition scheme [13] and hence the amount of data exchange is smaller than with the HPF code which basically uses onedimensional decomposition (parallelizing one loop only). Because of the different architectures and memory sizes, it is hard to compare the execution times of the T3E and IBM SP2.…”
Section: Performance Resultsmentioning
confidence: 99%
See 1 more Smart Citation
“…For comparison purposes, we also include the speedups of the same code running on an IBM SP2 machine using MPI in Figure 3. The MPI code uses a three-dimensional decomposition scheme [13] and hence the amount of data exchange is smaller than with the HPF code which basically uses onedimensional decomposition (parallelizing one loop only). Because of the different architectures and memory sizes, it is hard to compare the execution times of the T3E and IBM SP2.…”
Section: Performance Resultsmentioning
confidence: 99%
“…One is the sequential Fortran version called MAX3D developed by the second author of this paper for vector machines such as the CRAY C90. The second is the MPI (Message-passing Interface) [11,12] version implemented by Dr Marcus Wagner of IBM [13]. The third one is the Power Fortran Accelerator (PA) [14] version developed by the first author when he was an Air Force Office for Scientific Research (AFOSR) Faculty Research Fellow at the Wright Patterson Air Force Base.…”
Section: Introductionmentioning
confidence: 99%
“…In the early 2000's, the data processing rate had reached a value of one GFLOPS (Giga floating point operation per second, 10 9 FLOPS) and the size of distributed memory has approached a terabyte. From basic research result, the load balancing, communication minimization among processors, and efficient cash memory utilization were found to be paramount for high parallel numerical simulation efficiency [67]. The versatility of unstructured Navier-Stokes/Euler solvers has been demonstrated for aircraft design.…”
Section: Landmark Achievementsmentioning
confidence: 99%
“…The collective modeling and simulation tools span a range, including the asymptotic method, the frequency-domain method (or method of moments, MoM), the time-domain method, and the more recent hybrid technique [471,535,647,857]. Individually, they may be limited either by predictive accuracy or by practicality in application.…”
Section: Performancementioning
confidence: 99%
“…First, concurrent computing significantly reduces the wall-clock time required for data processing. Recent research efforts in porting CEM programs to multicomputers have recorded two orders of magnitude of speedup in the data processing rate [857].…”
Section: Performancementioning
confidence: 99%