I would like to evaluate memory performance of GPU kernels, specifically, amount of data moved across different memories and effective memory bandwidth. Any pointers on how to get this information?
Thanks!
Rajesh
Related topics
| Topic | Replies | Views | Activity | |
|---|---|---|---|---|
| analysis inside kernel | 2 | 1430 | July 2, 2012 | |
| Maximizing Unified Memory Performance in CUDA | 0 | 233 | August 21, 2022 | |
| benchmarks | 2 | 440 | June 10, 2019 | |
| What's a reasonable memory bandwidth performance to expect? My current maximum is only around 50 | 1 | 631 | July 27, 2010 | |
| GPU Memory how to find the GPU memory bandwidth | 10 | 17430 | June 23, 2007 | |
| CUDA: Memory performance, What is Global memory bandwidth | 2 | 6217 | November 2, 2011 | |
| Comparison for coalesced memory access? | 0 | 606 | October 6, 2015 | |
| Global Memory Bandwidth. | 0 | 815 | December 8, 2009 | |
| weird performance in GPU | 2 | 2701 | October 17, 2011 | |
| Calculating global memory bandwidth for GPUs with compute capability lower than 1.2 | 0 | 4440 | January 26, 2010 |