There isn’t any way to request WC memory with cudaHostRegister(). If you would like to see a change in CUDA, you can file a bug.
Related topics
| Topic | Replies | Views | Activity | |
|---|---|---|---|---|
| How to turn host memory into page-locked using CUDA driver API | 1 | 522 | June 10, 2017 | |
| problem with zero-copy write to write-combined memory. | 2 | 806 | September 20, 2014 | |
| Using cudaHostRegister() in CUDA 4.0 CUDA 4.0 | 16 | 30683 | January 25, 2018 | |
| Unable get over 512MB of page-locked memory with cudaHostRegister or cudaMallocHost... | 3 | 3167 | July 2, 2012 | |
| Using memory already page-locked by a driver? | 0 | 556 | October 13, 2016 | |
| Shared page-locked memory | 0 | 719 | June 9, 2010 | |
| How to pass two flags to cudaHostAlloc()? | 5 | 9371 | June 17, 2009 | |
| Already locked memory: making it known to CUDA - how ? | 2 | 937 | September 6, 2016 | |
| cudaHostRegister and cudaHostRegisterPortable, cudaHostRegisterDefault | 7 | 7250 | January 13, 2012 | |
| Does the page-lock memory by cudaHostRegister slow than cudaMallocHost? | 9 | 1104 | June 30, 2023 |