Atomics (GPU)
#565
Replies: 0 comments
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
{{ message }}
-
GPU Atomic operations
CUDA
SYCL
ROC/HIP
Scope of atomics
Memory model for atomics
relaxed, acq-rel, seq-cst and so forth
what actually works
local/shared memory
global memory
scale-up memory (nvLink, xeLink, etc)
host memory
rdma atomics to device memory
Beta Was this translation helpful? Give feedback.
All reactions