I inherited an image processing program. It uses cufftExecR2C followed by cufftExecC2R as an image processing filter. For some reason it works perfectly fine with one set of images but then fouls up with another set. It does not go into error mode. It just fouls up the data and produces total garbage. The images are the same format, size, everything, except they have different values. If I do an end run around this part of the algorithm and remove cufftExecR2C/cufftExecC2R altogether the final image product comes out blank. Why does cufftExecR2C/cufftExecC2R fail internally due to different values in the matrices that I provide as input? How can I debug this or get around this?
Related topics
| Topic | Replies | Views | Activity | |
|---|---|---|---|---|
| 1D cufft R2C--filter---C2R problem | 1 | 753 | October 6, 2011 | |
| 2D CUFFT problem | 1 | 692 | February 9, 2012 | |
| CUFFT on image | 2 | 2431 | April 29, 2015 | |
| weird cufft problem - works sometimes but not other times problem with cufftExecC2C working only spo | 1 | 558 | February 15, 2011 | |
| weird cufft problem - works sometimes but not other times problem with cufftExecC2C working only spo | 0 | 3693 | January 27, 2011 | |
| CUFFT - errors depend on size of 2-D array? Errors seen in FFT results - depending on the size of th | 1 | 3304 | November 12, 2008 | |
| CUFFT error? | 0 | 3385 | November 12, 2008 | |
| Problem with CUFFT R2C+C2R returning NaNs | 3 | 1614 | June 25, 2012 | |
| cufft functionality | 7 | 1213 | January 12, 2012 | |
| cufft 2d size problems | 5 | 3137 | March 7, 2010 |