nanding0701

I’m testing the roc_shmem using two nodes on frontier. I find that if I did MPI_Init in the code, and then do roc_shmem_init, I’ll get the below error msg. Assertion...

Does ROC_SHMEM have roc_shmem_##TNAME##_wait_util_any? I failed to find it.

Does ROC_SHMEM have a collective launch to avoid potential deadlocks? https://docs.nvidia.com/nvshmem/api/api/launch.html#nvshmemx-collective-launch

Hi, I wonder if ROC_SHMEM can support "put with signal API"(https://docs.nvidia.com/nvshmem/api/gen/api/signal.html#nvshmem-put-signal-nbi). It will speed up the inter-node GPU communication in my experience.

Hi there, How to build the library with only reverse offload? I failed to find the documentation. I want to try ROC_SHMEM on Crusher. However, Crusher doesn't have the InfiniBand...