Analyzers
Talk to fellow users of Intel Analyzer tools (Intel VTune™ Profiler, Intel Advisor)
Announcements
FPGA community forums and blogs on community.intel.com are migrating to the new Altera Community and are read-only. For urgent support needs during this transition, please visit the FPGA Design Resources page or contact an Altera Authorized Distributor.
5255 Discussions

Intel Advisor 2021.1 : What are the percentage numbers next to each module under module filter?

mikeitexpert
New Contributor II
2,341 Views

Intel Advisor 2021.1 :

 

What are the percentage numbers next to each module under module filter?

 

Please let me know.

 

mikeitexpert_1-1624097624889.png

 

 

 

 

0 Kudos
1 Solution
RashmiP_Intel
Moderator
2,214 Views

Hi,


Intel® VTune™ Profiler optimizes application performance, system performance, and system configuration for HPC, cloud, IoT, media, storage, and more.

So you can also use vtune for this question.

You can refer to https://software.intel.com/content/www/us/en/develop/documentation/vtune-cookbook/top/tuning-recipes/openmp-imbalance-and-scheduling-overhead.html this cookbook.


Regards,

Rashmi


View solution in original post

10 Replies
Ruslan_M_Intel
Employee
2,306 Views

It's amount of time spent within corresponding module

0 Kudos
mikeitexpert
New Contributor II
2,287 Views

Thanks for the input ... 

Do you think it is a sign for over parallelization? If so, how would I be able to better  track down the problem? I speculate it has too do with forcing parallelizing using #pragma omp parallel for ... so I may consider using if clauses just to avoid unnecessary parallel execution. However, I am not sure how to verify my guess ... 

 

Please let me know what you think ... 

 

Regards

 

0 Kudos
Ruslan_M_Intel
Employee
2,281 Views

Yes, the overhead you faced with can be caused by the reason you are talking about. You can try playing with OMP_NUM_THREADS env var (e.g. 1, 4 and

mikeitexpert
New Contributor II
2,207 Views

Ruslan,

Thanks for the comment I lowered the number of threads to 1 and it turns out it reduces libiomp5md.dll time. Though I am not sure whether it is a good sign or bad sign ??? I am just trying to find whether I have unnecessary parallelizing or not.

 

mikeitexpert_3-1624629844742.png

 

 

 

 

0 Kudos
mikeitexpert
New Contributor II
2,269 Views

Is there any way to force IA (Intel Advisor) use omp debug dlls just to better find out ? I am not sure how I can use debug dlls....

I have tried below but no luck .... 

 

mikeitexpert_0-1624509748239.png

 

0 Kudos
Ruslan_M_Intel
Employee
2,260 Views

Unfortunately, no. Advisor doesn't do anything like that. Could you tell what you're going to achieve doing that?

0 Kudos
RashmiP_Intel
Moderator
2,215 Views

Hi,


Intel® VTune™ Profiler optimizes application performance, system performance, and system configuration for HPC, cloud, IoT, media, storage, and more.

So you can also use vtune for this question.

You can refer to https://software.intel.com/content/www/us/en/develop/documentation/vtune-cookbook/top/tuning-recipes/openmp-imbalance-and-scheduling-overhead.html this cookbook.


Regards,

Rashmi


mikeitexpert
New Contributor II
2,161 Views

Feel free to close the thread ...

Thank you

0 Kudos
RashmiP_Intel
Moderator
2,148 Views

Hi,


Thanks for accepting our solution. If you need any additional information, please submit a new question as this thread will no longer be monitored.


Regards,

Rashmi


0 Kudos
Reply