[PR #482] Merge math.hpp, replace Kpack4 with torch::kFloat4_e2m1fn_x2 #481

New issue

Open

opened 2026-10-26 21:18:22 +00:00 by bsakaguchi · 0 comments

bsakaguchi commented

2026-10-26 21:18:22 +00:00

Owner

📋 Pull Request Information

Original PR: https://github.com/deepseek-ai/DeepGEMM/pull/482
Author: @john-kernel
Created: 10/24/2026
Status: 🔄 Open

Base: main ← Head: main

📝 Commits (2)

f7e0baa Merge math.hpp, replace Kpack4 with torch::kFloat4_e2m1fn_x2
83625cc revert dtype conversion from merge. Return kpack4 constexpr

📊 Changes

15 files changed (+44 additions, -42 deletions)

View changed files

📝 csrc/apis/locality_domain.hpp (+1 -1)
📝 csrc/jit_kernels/heuristics/config.hpp (+1 -1)
📝 csrc/jit_kernels/heuristics/mega_gate.hpp (+1 -1)
📝 csrc/jit_kernels/heuristics/mega_moe.hpp (+1 -1)
📝 csrc/jit_kernels/impls/runtime_utils.hpp (+1 -1)
📝 csrc/jit_kernels/impls/sm100_bf16_gemm.hpp (+1 -1)
📝 csrc/jit_kernels/impls/sm100_bmk_bnk_mn.hpp (+1 -1)
📝 csrc/jit_kernels/impls/sm100_fp8_fp4_gemm_1d1d.hpp (+1 -1)
📝 csrc/jit_kernels/impls/sm100_tf32_hc_prenorm_gemm.hpp (+1 -1)
📝 csrc/jit_kernels/impls/sm90_bmk_bnk_mn.hpp (+1 -1)
📝 csrc/jit_kernels/impls/sm90_tf32_hc_prenorm_gemm.hpp (+1 -1)
📝 csrc/jit_kernels/impls/smxx_layout.hpp (+1 -1)
📝 csrc/utils/layout.hpp (+1 -1)
➖ csrc/utils/math.hpp (+0 -29)
📝 deep_gemm/include/deep_gemm/common/math.cuh (+31 -0)

📄 Description

Fulfilled todo's listed in original math.hpp file.

_{🔄 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/482 **Author:** [@john-kernel](https://github.com/john-kernel) **Created:** 10/24/2026 **Status:** 🔄 Open **Base:** `main` ← **Head:** `main` --- ### 📝 Commits (2) - [`f7e0baa`](https://github.com/deepseek-ai/DeepGEMM/commit/f7e0baa5d454c6f3f9ae93adab22e90520eb1586) Merge math.hpp, replace Kpack4 with torch::kFloat4_e2m1fn_x2 - [`83625cc`](https://github.com/deepseek-ai/DeepGEMM/commit/83625cc52dd3b1867231156e6f135e40f4263db2) revert dtype conversion from merge. Return kpack4 constexpr ### 📊 Changes **15 files changed** (+44 additions, -42 deletions) <details> <summary>View changed files</summary> 📝 `csrc/apis/locality_domain.hpp` (+1 -1) 📝 `csrc/jit_kernels/heuristics/config.hpp` (+1 -1) 📝 `csrc/jit_kernels/heuristics/mega_gate.hpp` (+1 -1) 📝 `csrc/jit_kernels/heuristics/mega_moe.hpp` (+1 -1) 📝 `csrc/jit_kernels/impls/runtime_utils.hpp` (+1 -1) 📝 `csrc/jit_kernels/impls/sm100_bf16_gemm.hpp` (+1 -1) 📝 `csrc/jit_kernels/impls/sm100_bmk_bnk_mn.hpp` (+1 -1) 📝 `csrc/jit_kernels/impls/sm100_fp8_fp4_gemm_1d1d.hpp` (+1 -1) 📝 `csrc/jit_kernels/impls/sm100_tf32_hc_prenorm_gemm.hpp` (+1 -1) 📝 `csrc/jit_kernels/impls/sm90_bmk_bnk_mn.hpp` (+1 -1) 📝 `csrc/jit_kernels/impls/sm90_tf32_hc_prenorm_gemm.hpp` (+1 -1) 📝 `csrc/jit_kernels/impls/smxx_layout.hpp` (+1 -1) 📝 `csrc/utils/layout.hpp` (+1 -1) ➖ `csrc/utils/math.hpp` (+0 -29) 📝 `deep_gemm/include/deep_gemm/common/math.cuh` (+31 -0) </details> ### 📄 Description Fulfilled todo's listed in original math.hpp file. --- <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 21:18:22 +00:00

No labels

pull-request

No milestone

No project

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/DeepGEMM#481

No description provided.

Rows
Columns