Quantcast
Channel: Intel® Software - Intel® Integrated Performance Primitives
Viewing all articles
Browse latest Browse all 1489

omp pragma different time

$
0
0

Hi,

I'm  working  on a complicated  project, that requires a parallel  calculations  in  order to achieve  good time performance.

Our  company  bought  for  this purpose  intel xeon platinum 8168  processor (96 CORES - name 96).  Also, we have a  computer  with intel core i9 7960x processor(16 CORES - name 16). 

I'm  using  "omp  pragma  for"  directive,  as all  calculations  happen in  FOR  loops. And at this point i got strange results.

I'm running  my code  on  16  PC,  with number of iterations in FOR loop  less than 16. That's mean, that number of threads & number of CORES  that are  used  less than 16. At this point, I got almost same time results(I mean, 5 iterations, 10 iterations, and 15 iterations  complete with almost same time). And this is correct, since NOT ALL CPU  power was  used.

At this point, I try  and run SAME  code  on  96. And I see strange time performance results. If I run 40 iterations(see 40 threads), the time is almost twice against 1 iteration. And if I run 90 iterations(still, NOT full power!!),  time increase almost 4 times. 

My questions is, does it have some issue to this specific processor (intel xeon platinum 8168  processor)  working with IPP libraries ? 

What could be the possible reason for such a time complexity increase?  I am aware about dynamic memory allocations , we have some,  and time needed to create large number of threads, but still  that's seem  not the real reason.

Thanks   

 


Viewing all articles
Browse latest Browse all 1489

Trending Articles



<script src="https://jsc.adskeeper.com/r/s/rssing.com.1596347.js" async> </script>