[PR #395] Add situ activation #413

New issue

Open

opened 2026-10-26 21:17:31 +00:00 by bsakaguchi · 0 comments

bsakaguchi commented

2026-10-26 21:17:31 +00:00

Owner

📋 Pull Request Information

Original PR: https://github.com/deepseek-ai/DeepGEMM/pull/395
Author: @ZJY0516
Created: 7/29/2026
Status: 🔄 Open

Base: nv_dev ← Head: situ-activation

📝 Commits (3)

f5a7642 init
51655f2 Merge branch 'main' into situ-activation
bf3aa9f Refactor SITU Mega MoE into separate kernels

📊 Changes

5 files changed (+1924 additions, -38 deletions)

View changed files

📝 csrc/apis/mega.hpp (+33 -19)
➕ csrc/jit_kernels/impls/sm100_fp8_fp4_mega_moe_situ.hpp (+327 -0)
➕ deep_gemm/include/deep_gemm/impls/sm100_fp8_fp4_mega_moe_situ.cuh (+1479 -0)
📝 deep_gemm/mega/__init__.py (+6 -4)
📝 tests/test_mega_moe.py (+79 -15)

📄 Description

Test

DG_JIT_CACHE_DIR="$PWD/.cache/deep_gemm" \
    python tests/test_mega_moe.py \
      --num-processes 1 \
      --num-max-tokens-per-rank 128 \
      --num-tokens 32 \
      --num-experts 8 \
      --num-topk 1 \
      --num-shared-experts 0 \
      --activation situ \
      --activation-clamp inf

_{🔄 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/395 **Author:** [@ZJY0516](https://github.com/ZJY0516) **Created:** 7/29/2026 **Status:** 🔄 Open **Base:** `nv_dev` ← **Head:** `situ-activation` --- ### 📝 Commits (3) - [`f5a7642`](https://github.com/deepseek-ai/DeepGEMM/commit/f5a76426fa084087169693fd0cd815223576d6e9) init - [`51655f2`](https://github.com/deepseek-ai/DeepGEMM/commit/51655f217b8b3b3875e59677bb5e8901f235826d) Merge branch 'main' into situ-activation - [`bf3aa9f`](https://github.com/deepseek-ai/DeepGEMM/commit/bf3aa9f2b1fd45cfff1329c723c1f6c45a1a6fde) Refactor SITU Mega MoE into separate kernels ### 📊 Changes **5 files changed** (+1924 additions, -38 deletions) <details> <summary>View changed files</summary> 📝 `csrc/apis/mega.hpp` (+33 -19) ➕ `csrc/jit_kernels/impls/sm100_fp8_fp4_mega_moe_situ.hpp` (+327 -0) ➕ `deep_gemm/include/deep_gemm/impls/sm100_fp8_fp4_mega_moe_situ.cuh` (+1479 -0) 📝 `deep_gemm/mega/__init__.py` (+6 -4) 📝 `tests/test_mega_moe.py` (+79 -15) </details> ### 📄 Description ## Test ``` DG_JIT_CACHE_DIR="$PWD/.cache/deep_gemm" \ python tests/test_mega_moe.py \ --num-processes 1 \ --num-max-tokens-per-rank 128 \ --num-tokens 32 \ --num-experts 8 \ --num-topk 1 \ --num-shared-experts 0 \ --activation situ \ --activation-clamp inf ``` --- <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 21:17:31 +00:00

bsakaguchi referenced this issue

2026-10-26 21:17:39 +00:00

[PR #413] [CLOSED] Fix integer environment parsing #426