mirror of
https://github.com/deepseek-ai/DeepEP.git
synced 2026-11-05 00:47:32 +00:00
[GH-ISSUE #115] 双节点3+9卡 RTX3090可以用不? #104
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/DeepEP#104
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 @liudun20150508 on GitHub (Apr 9, 2025).
Original GitHub issue: https://github.com/deepseek-ai/DeepEP/issues/115
我用的2个节点 1个节点是3个RTX3090 1个节点是8个RTX3090 如果用这个深度EP 可以实现微调不?需要我改哪些地方?
@sphish commented on GitHub (Apr 9, 2025):
不行,游戏卡不支持 deepep 必须的 GPU Direct RDMA。
@liudun20150508 commented on GitHub (Apr 10, 2025):
GPU Direct RDMA 不是只是和网卡有关系吗?怎么和游戏卡有关系?这个不是网卡决定的吗?ifiniband?
@sphish commented on GitHub (Apr 10, 2025):
GPU Direct RDMA 需要 GPU 提供地址映射,NV 在驱动层面禁止了游戏卡使用这个功能。
@liudun20150508 commented on GitHub (Apr 10, 2025):
GPU Direct RDMA实际上就是指NCCL吧?pytorch分布式是内置了NCCL吧?不需要显示去调用?按照你这样说RTX3090没法做pytorch的分布式?或者说根本没法使用NCCL?因为没法用GPU Direct RDMA?
@liudun20150508 commented on GitHub (Apr 10, 2025):
关键这个EP用的也不是NCCL 也不是pytorch的分布式啊。这个是比较细的 感觉要用PTX去改写的。就是NVSHEM这个东西。和NCCL或者pytroch分布式不一样吧?