|
[Sponsors] |
October 25, 2000, 07:44 |
speed up ratio at parallel processing
|
#1 |
Guest
Posts: n/a
|
I'm studying about parallel processing at distributed computing network using some different kinds of computers like PC & workstation. However, I'm concerning about the speed up ratio equation suitable for that system. As far as I know, the old speed up ratio equation(Sp = T1 / Tp) can't give a correct ratio because it needs same computers to use that eqn. So, I'm wonder whether or not there's another equation for that system. Anybody who know about it, please tell me....
|
|
October 25, 2000, 10:57 |
Re: speed up ratio at parallel processing
|
#2 |
Guest
Posts: n/a
|
Yes, You are right. I met the same qustion when I did parallel computation by using PC clusters. According to original formula, we can assume that a1 represent the serial part which can not be parallized and if we neglect the communication time,
the maximum speedup ratio = 1/(a1+(1-a1)/n) , where n represent the process number . If a1 = 0.1 , then the speedup ratio < 10 , so this formula is not rigorous , I saw another formula and I am looking for it . |
|
Thread Tools | Search this Thread |
Display Modes | |
|
|
Similar Threads | ||||
Thread | Thread Starter | Forum | Replies | Last Post |
multiphase solver - parallel processing - GAMG | thibault_pringuey | OpenFOAM Programming & Development | 2 | August 27, 2013 23:03 |
pressure eq. "converges" after few time steps | maddalena | OpenFOAM Running, Solving & CFD | 69 | July 21, 2011 08:42 |
HP MPI warning...Distributed parallel processing | Peter | CFX | 10 | May 14, 2011 07:17 |
Hardware Recommendation for Parallel Processing | Brian Bian | CFX | 2 | February 7, 2006 18:27 |
About parallel processing in Linux | tuks | CFX | 10 | August 8, 2005 09:22 |