Revision history [back]
 | 1 | initial version |
I have some experience with PSS/E on this topic. Based on my experience, PSS/E does not improve performance just by having multiple cores on your machine unless you explicitly have parallel processing scripts which split the job into multiple cores. If you have your scripts configured to run your analysis in parallel, you are way better off by having multiple cores. With multiple cores, most analysis which can be performed in parallel (eg. contingency analysis), speed up almost proportional to the no of cores used (minus some overhead). With a 56 physical core machine, we were able to increase our dynamics simulation speed by about 50 times compared to regular i7 laptops.
If your process is not parallel ready (eg. cascading analysis when you wait for the results of one analysis to start another phase, etc.), a stronger single core may make sense compared to typical single cores.
| | 2 | No.2 Revision |
I have some experience with PSS/E on this topic. Based on my experience, PSS/E does not improve performance just by having multiple cores on your machine unless you explicitly have parallel processing scripts which split the job into multiple cores. If you have your scripts configured to run your analysis in parallel, you are way better off by having multiple cores. With multiple cores, most analysis which can be performed in parallel (eg. dynamics contingency analysis), speed up almost proportional to the no of cores used (minus some overhead). With a 56 physical core machine, we were able to increase our dynamics simulation speed by about 50 times compared to regular i7 laptops.
If your process is not parallel ready (eg. cascading analysis when you wait for the results of one analysis to start another phase, etc.), a stronger single core may make sense compared to typical single cores.
