[SWDEV-497665] Blocked `cudaMemcpyAsync` race condition by synchronizing#1447

Open

corey-derochie-amd wants to merge 2 commits intoROCm:developfrom corey-derochie-amd:SWDEV-497665_fix-async-memcpy-crash

+2

Commits on Dec 4, 2024

Switched calls to cudaMemcpyAsync to be cudaMemcpy in ncclTransportP2pSetup to avoid race condition with cudaIpcOpenMemHandle inside p2p connect. See ncclP2pImportShareableBuffer.
corey-derochie-amd
committed

Commits on Dec 9, 2024

Moved synchronize outside of the loop, as it isn't necessary to sync between every iteration of the loop.
corey-derochie-amd
committed