Skip to content

test(cuda.core): synchronize IPC buffer initialization #2657

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Open
Andy-Jost wants to merge 1 commit into
base: main
Choose a base branch
from
+2 −0
Open

test(cuda.core): synchronize IPC buffer initialization #2657

Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
2 changes: 2 additions & 0 deletions cuda_core/tests/memory_ipc/test_peer_access.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -82,6 +82,8 @@ def test_main(self, ipc_mempool_device_x2, grant_access_in_parent):
buffer = mr.allocate(NBYTES, stream=dev1.default_stream)

@rwgk rwgk Aug 19, 2026

Copy link
Copy Markdown
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

codex gpt-5.6-sol ultra found:

This allocates asynchronously on dev1.default_stream, while PatternGen(dev1, ...) creates a separate nonblocking stream and initializes the buffer there. The later dev1.sync() waits for both streams, but does not establish allocation-before-copy ordering. CUDA defines that cross-stream access as undefined unless explicitly ordered. CUDA documentation (https://docs.nvidia.com/cuda/cuda-programming-guide/04-special-topics/stream-ordered-memory-allocation.html).

I’d use one stream throughout:

        stream = dev1.default_stream
        buffer = mr.allocate(NBYTES, stream=stream)
        pgen = PatternGen(dev1, NBYTES, stream=stream)
        pgen.fill_buffer(buffer, seed=False)
        stream.sync()

That also avoids a device-wide barrier.

pgen = PatternGen(dev1, NBYTES)
pgen.fill_buffer(buffer, seed=False)
# IPC import does not carry the exporting process's stream ordering.
dev1.sync()

# Spawn child process
process = mp.Process(target=self.child_main, args=(mr, buffer))
Expand Down
Loading