topic processor peak vs. effective performance on parallel platforms in Intel® MPI Library

processor peak vs. effective performance on parallel platforms question

notahoo — Wed, 07 Mar 2012 23:22:18 GMT

Hi,

Knowing the peak floating point operations rate of a processor, is there any rule of thumb for predicting the 'effective' FLOPs? In another words, how many percent of the peak FLOPs, in general, is the effective FLOPs?

Thanks for any insight.

processor peak vs. effective performance on parallel platforms

James_T_Intel — Thu, 08 Mar 2012 16:30:14 GMT

Hi notahoo,

Actual performance will be heavily dependent on the application. I would recommend reading the article at http://software.intel.com/en-us/articles/estimating-flops-using-event-based-sampling-ebs/. This should give you some ideas of how you can estimate your application's performance.

Sincerely,
James Tullos
Technical Consulting Engineer
Intel Cluster Tools

processor peak vs. effective performance on parallel platforms

TimP — Thu, 08 Mar 2012 21:02:39 GMT

In the Top500 supercomputer ratings, an "efficiency" is quoted as the ratio of actual performance to the peak flops rating. This rating can be achieved only where it is possible to use the vendor's optimized BLAS (MKL dgemm, in the case of Intel clusters) and adjust the problem size to maximize efficiency, and the memory bandwidth is not a significant limitation, as it would be for many real applications.