deepseek-ai/DeepEP
1
0
Fork 0
mirror of https://github.com/deepseek-ai/DeepEP.git synced 2026-11-05 00:47:32 +00:00
Code Issues 308 Projects Packages Wiki Activity

[GH-ISSUE #467] Questions About AR Support for Low-latency #301

New issue
Open
opened 2026-10-26 09:58:15 +00:00 by bsakaguchi · 1 comment
bsakaguchi commented 2026-10-26 09:58:15 +00:00
Owner
Copy link

Originally created by @michaelchen1996 on GitHub (Oct 27, 2025).
Original GitHub issue: https://github.com/deepseek-ai/DeepEP/issues/467

Commit: 2d0cf41dd1900b105d74cb071f4cac35e3fb6f47​
Low latency kernels use rdma atomic to support AR.​

This commit replaces two functions:​

  • nvshmemi_ibgda_rma_p​
  • nvshmemi_ibgda_poll_recv​

with a new function: nvshmemi_ibgda_amo_nonfetch_add
​
From code inspection, nvshmemi_ibgda_rma_p appears to use RDMA Write with Immediate Data.​
Does RDMA Write with IMM not support AR? Why we have to use nvshmemi_ibgda_amo_nonfetch_add to add AR support for low-latency ?​
Thank you!

Originally created by @michaelchen1996 on GitHub (Oct 27, 2025). Original GitHub issue: https://github.com/deepseek-ai/DeepEP/issues/467 Commit: [2d0cf41dd1900b105d74cb071f4cac35e3fb6f47](https://github.com/nvshmem/nvshmem/commit/2d0cf41dd1900b105d74cb071f4cac35e3fb6f47)​ Low latency kernels use rdma atomic to support AR.​ This commit replaces two functions:​ - `nvshmemi_ibgda_rma_p​` - `nvshmemi_ibgda_poll_recv​` with a new function: `nvshmemi_ibgda_amo_nonfetch_add` ​ From code inspection, `nvshmemi_ibgda_rma_p` appears to use **RDMA Write with Immediate Data**.​ Does RDMA Write with IMM not support AR? Why we have to use `nvshmemi_ibgda_amo_nonfetch_add` to add AR support for low-latency ?​ Thank you!
bsakaguchi commented 2026-10-26 09:58:17 +00:00
Author
Owner
Copy link

@michaelchen1996 commented on GitHub (Oct 27, 2025):

Additionally, may I ask that if this is because NVSHMEM does not have a write with imm operation, why is nvshmemi_ibgda_rma_p with the imm parameter still retained?

<!-- gh-comment-id:3450818854 --> @michaelchen1996 commented on GitHub (Oct 27, 2025): Additionally, may I ask that if this is because NVSHMEM does not have a write with imm operation, why is `nvshmemi_ibgda_rma_p` with the `imm` parameter still retained?
Sign in to join this conversation.
No Branch/Tag specified
Branches Tags
main
hybrid-ep
chore/remove-clang-format-yapf
v2.5
nvDev
try-fix-cu128
mori-ep
antgroup-opt
compat-old-torch
add-linter
eager-rdma
v1.2.1
Labels
Clear labels   
pull-request
Mirrored from GitHub Pull Request

No labels
pull-request
Milestone
Clear milestone
No items
No milestone
Projects
Clear projects
No items
No project
Assignees
Clear assignees
No assignees
1 participant
Notifications
Due date
The due date is invalid or out of range. Please use the format "yyyy-mm-dd".

No due date set.

Dependencies

No dependencies set.

Reference
deepseek-ai/DeepEP#301
No description provided.