Just to close this off, in the end I did identify a race condition that presumably never caused trouble on older hardware/software somehow. It did require compiling with CUDA 10 and running cuda-memcheck there (although the code causes more problems with CUDA 10 it runs far enough to get racecheck through the relevant errors). I do not know why cuda-memcheck broke with CUDA 8 but not CUDA 10, but there we go. Thanks for the help!
Related topics
| Topic | Replies | Views | Activity | |
|---|---|---|---|---|
| initcheck and malloc/free don't get along | 1 | 656 | May 19, 2016 | |
| Cuda8 samples crash but work with memcheck | 3 | 455 | October 18, 2021 | |
| Potential Bug, cuda-memcheck can someone verify? Program crashing on GPU initialisation with cuda-me | 11 | 3473 | April 24, 2020 | |
| cuda-memcheck errors yesterday no, today yes | 6 | 1488 | May 2, 2012 | |
| Segmentation fault Problem from 1.0 to 2.0 SUSE Enterprise | 6 | 3726 | November 16, 2008 | |
| Unable to use cuda-memcheck --tool racecheck | 3 | 991 | October 12, 2021 | |
| cuda-memcheck reports errors when unified memory is allocated | 0 | 1730 | October 27, 2015 | |
| a question about cuda-memcheck | 4 | 703 | April 16, 2018 | |
| My GPU code works well in cuda-memcheck model but crashes in normal running model | 2 | 703 | October 9, 2015 | |
| CUDA Runtime Error Only With Memcheck | 4 | 1042 | February 27, 2019 |