I am trying to intercept cudaMemcpy calls from the pytorch library for analysis. I noticed NVIDIA has a cuHook example in the CUDA toolkit samples. However that example requires one to modify the source code of the application itself which I cannot do in this case. So is there a way to write a hook to intercept CUDA calls without modifying the application source code?
Related topics
| Topic | Replies | Views | Activity | |
|---|---|---|---|---|
| cuda.h & cuda_runtime_api.h interception layers with CuHook | 0 | 706 | July 15, 2016 | |
| Intercept cudaMallocPitch | 3 | 1333 | August 27, 2019 | |
| cuMemAlloc hook does not work on cudaMalloc | 4 | 670 | October 26, 2022 | |
| Can you intercept CUDA calls ? | 2 | 2333 | April 1, 2009 | |
| Interrupting a CUDA call | 1 | 489 | November 5, 2018 | |
| How to hook cuda api in the tensorflow? | 3 | 1205 | November 9, 2021 | |
| Layout of source code | 3 | 546 | April 29, 2011 | |
| Information about Cuda Internal APIs | 5 | 129 | October 30, 2024 | |
| How to "wrap" function calls for running in CUDA ? | 1 | 2442 | November 24, 2007 | |
| Change source code? How exactly... | 6 | 2462 | October 3, 2008 |