My application seems to run file for small problem size, but always segfault/out of bounds when the problem size gets significantly large. What tools would you recommend to check which memory the GPU runs out? Whether it’s per-thread, shared or global?
Related topics
| Topic | Replies | Views | Activity | |
|---|---|---|---|---|
| How to check available global memory size? | 4 | 3026 | March 18, 2010 | |
| Should out of bounds shared memory accesses be errors? | 5 | 7113 | April 6, 2009 | |
| Detect how much global memory is available? | 2 | 766 | August 21, 2010 | |
| How to find out how much global memory is being used? | 1 | 3278 | January 4, 2010 | |
| Generic memory management & optimization | 6 | 1342 | November 5, 2015 | |
| global memory consumed | 1 | 2499 | October 14, 2009 | |
| How to find out GPU memory usage (from kernel not host) | 2 | 823 | October 29, 2017 | |
| CUDA and GPU memory How to get CUDA to exit cleanly when a routine demands too much memory | 2 | 7546 | February 11, 2010 | |
| large thread question | 2 | 1149 | June 5, 2009 | |
| Runtime memory leaks? How do you clear them? | 3 | 5066 | October 22, 2008 |