- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
Hi,
I am trying to parallelise my code, but I get an access violation error, and my program crashes.
here is my code snippet :
#pragma omp parallel for for (i=0 ; i{
{
/* Using Intel's IPPi IDCT */ippiDCT8x8Inv_16s_C1I(DiffMB);
}
}
I am using Intel compiler and Intel Performance Primitives ver 4.1.Please help,
Elad
Link Copied
2 Replies
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
If your IPP function itself is threaded,that threading would have to be disabled. If IPP were OpenMP threaded, it would be taken care of automatically, but I don't believe this is the case. If your IPP docs don't cover this, you might ask in the IPP forum.
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
I note that the IPP dynamic libraries are claimed to be thread safe, not the static ones.
ippiDCT functions are not listed as being OpenMP enabled. In case of doubt about library thread safety, it may be possible to run with Intel Thread Checker, to see if a thread safety problem is found.
![](/skins/images/C4F7392A5C68ACA60290C9576F107044/responsive_peak/images/icon_anonymous_message.png)
Reply
Topic Options
- Subscribe to RSS Feed
- Mark Topic as New
- Mark Topic as Read
- Float this Topic for Current User
- Bookmark
- Subscribe
- Printer Friendly Page