mirror of
https://github.com/deepseek-ai/DeepGEMM.git
synced 2026-11-05 00:47:26 +00:00
[GH-ISSUE #334] [Feature Request] Roadmap for Ascend NPU support in MegaMoE? #164
Labels
No labels
pull-request
No milestone
No project
No assignees
1 participant
Notifications
Due date
No due date set.
Dependencies
No dependencies set.
Reference
deepseek-ai/DeepGEMM#164
Loading…
Add table
Add a link
Reference in a new issue
No description provided.
Delete branch "%!s()"
Deleting a branch is permanent. Although the deleted branch may continue to exist for a short time before it actually gets removed, it CANNOT be undone in most cases. Continue?
Originally created by @leo-lilinxiao on GitHub (May 13, 2026).
Original GitHub issue: https://github.com/deepseek-ai/DeepGEMM/issues/334
Hi DeepGEMM team,
Thanks for the great work.
The DeepSeek-V4 report mentions that the fine-grained EP overlap scheme was validated on both NVIDIA GPUs and Huawei Ascend NPUs. However, the current DeepGEMM/MegaMoE implementation seems CUDA/NVIDIA-focused.
Is there any plan to support Huawei Ascend NPUs, especially Ascend A3/A5, for MegaMoE? If so, will the Ascend backend be open-sourced in this repo or somewhere else?
Thanks!
@zheanxu commented on GitHub (May 13, 2026):
Thank you for your question. We appreciate your interest in DeepGEMM. However, we are unable to discuss future plans or support for other hardware at this time. Please stay tuned to our official channels for any announcements.