- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
Hi--
I am using the MKL 'vsldConvExec' routines to perform n-dimensional convolution of arrays. I was under the impression, from the programmer's guide, that these libraries were easily multithreaded on an OpenMP-capable machine by just setting the number of threads using the OPM pragmas or setting an environment variable prior to execution. Is this the case? I am running my code on an SGI Altix machine.
Is there an easy way to get these MKL convolution routines to run on multiple processors? Or do I have to actually implement scalable convolutions?
Thanks,
Rob
I am using the MKL 'vsldConvExec' routines to perform n-dimensional convolution of arrays. I was under the impression, from the programmer's guide, that these libraries were easily multithreaded on an OpenMP-capable machine by just setting the number of threads using the OPM pragmas or setting an environment variable prior to execution. Is this the case? I am running my code on an SGI Altix machine.
Is there an easy way to get these MKL convolution routines to run on multiple processors? Or do I have to actually implement scalable convolutions?
Thanks,
Rob
Link Copied
1 Reply
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
Parallelization feature isn`t added to VSL functions. They will not threaded if you use non-sequential library (mkl_intel_thread.lib for example) in your project.
Andrey
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