- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
Hello,
I am computing 1D forward FFT's using DftiComputeForward and am not seeing a change in performance when I run on mulit-core machines. I have tried setting OMP_NUM_THREADS=numProcs. Can someone please tell me what I need to do to enable mulitprocessor scaling?
Thank you,
Bob
I am computing 1D forward FFT's using DftiComputeForward and am not seeing a change in performance when I run on mulit-core machines. I have tried setting OMP_NUM_THREADS=numProcs. Can someone please tell me what I need to do to enable mulitprocessor scaling?
Thank you,
Bob
Link Copied
1 Reply
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
I actually found the answer to my question. There is no parralelization when computing a single fft at a time. However, multiple FFT's can be automatically done in parallel.

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