[PR #424] docs: note Mega MoE requires PyTorch >= 2.9 in Requirements #436

New issue

Open

opened 2026-10-26 21:17:45 +00:00 by bsakaguchi · 0 comments

bsakaguchi commented

2026-10-26 21:17:45 +00:00

Owner

📋 Pull Request Information

Original PR: https://github.com/deepseek-ai/DeepGEMM/pull/424
Author: @pengfaming
Created: 8/28/2026
Status: 🔄 Open

Base: main ← Head: fix/readme-mega-moe-torch-version

📝 Commits (2)

9619e25 docs: fix stale test file link in README
2a2edb9 docs: note Mega MoE requires PyTorch >= 2.9 in Requirements

📊 Changes

1 file changed (+2 additions, -2 deletions)

View changed files

📝 README.md (+2 -2)

📄 Description

What

The Requirements section lists PyTorch 2.1 or higher, but the Mega MoE API (get_symm_buffer_for_mega_moe) relies on torch.distributed._symmetric_memory, which requires PyTorch >= 2.9. This requirement is currently only mentioned in a comment inside the Mega MoE example. Add a note to the Requirements list so readers don't assume 2.1 covers all features.

Why

Clarifies the real version requirement for Mega MoE up front.

How I checked

deep_gemm/mega/__init__.py imports torch.distributed._symmetric_memory.
The Mega MoE example in README already notes # NOTES: requires PyTorch >= 2.9.

_{🔄 This issue represents a GitHub Pull Request. It cannot be merged through Gitea due to API limitations.}

## 📋 Pull Request Information **Original PR:** https://github.com/deepseek-ai/DeepGEMM/pull/424 **Author:** [@pengfaming](https://github.com/pengfaming) **Created:** 8/28/2026 **Status:** 🔄 Open **Base:** `main` ← **Head:** `fix/readme-mega-moe-torch-version` --- ### 📝 Commits (2) - [`9619e25`](https://github.com/deepseek-ai/DeepGEMM/commit/9619e256c9f721c8e08df6545190ff097bc0fe5f) docs: fix stale test file link in README - [`2a2edb9`](https://github.com/deepseek-ai/DeepGEMM/commit/2a2edb938fcd5683fb022573d8bb2403e7442a66) docs: note Mega MoE requires PyTorch >= 2.9 in Requirements ### 📊 Changes **1 file changed** (+2 additions, -2 deletions) <details> <summary>View changed files</summary> 📝 `README.md` (+2 -2) </details> ### 📄 Description ### What The Requirements section lists `PyTorch 2.1 or higher`, but the Mega MoE API (`get_symm_buffer_for_mega_moe`) relies on `torch.distributed._symmetric_memory`, which requires PyTorch >= 2.9. This requirement is currently only mentioned in a comment inside the Mega MoE example. Add a note to the Requirements list so readers don't assume 2.1 covers all features. ### Why Clarifies the real version requirement for Mega MoE up front. ### How I checked - `deep_gemm/mega/__init__.py` imports `torch.distributed._symmetric_memory`. - The Mega MoE example in README already notes `# NOTES: requires PyTorch >= 2.9`. --- <sub>🔄 This issue represents a GitHub Pull Request. It cannot be merged through Gitea due to API limitations.</sub>

bsakaguchi added the

pull-request

label

2026-10-26 21:17:45 +00:00

bsakaguchi referenced this issue

2026-10-26 21:17:57 +00:00

[PR #436] Improve SM100 MX MQA logits clean-output performance #444