---
type: "raw_signal"
id: "deepgemm-ascend-deepgemm-ascend-clean-and-efficient-matrix-4c68df"
slug: "deepgemm-ascend-deepgemm-ascend-clean-and-efficient-matrix-4c68df"
title: "DeepGEMM-Ascend: DeepGEMM-Ascend: clean and efficient matrix multiplication kernel library for Huawei Ascend…"
description: "DeepSeek published DeepGEMM-Ascend, an open-source port of its matrix-kernel library targeting Huawei Ascend NPUs."
retrieval_nugget: "DeepSeek published DeepGEMM-Ascend, an open-source port of its matrix-kernel library targeting Huawei Ascend NPUs."
observed_at: "2026-09-30"
record_date: "2026-09-30"
date_kind: "observed_at"
why_it_matters: "Source: https://github.com/deepseek-ai/DeepGEMM-Ascend"
novelty: "notable"
verification_level: "source-linked"
signal_type: "release"
source_platform: "github.com"
topics: []
entities: []
related_patterns: []
source_url: "https://github.com/deepseek-ai/DeepGEMM-Ascend"
source_urls: ["https://github.com/deepseek-ai/DeepGEMM-Ascend"]
schema_version: "newruntime-agent-readable-v0.2"
stable_id: "signal:deepgemm-ascend-deepgemm-ascend-clean-and-efficient-matrix-4c68df"
status: "published"
visuals: []
editor: {"role":"editor","name":"Andrey Reshetnikov","profile":"https://newruntime.com/owner-profile.md","profile_json":"https://newruntime.com/owner-profile.json","same_as":["https://www.linkedin.com/in/reshetnikov1/","https://t.me/lapetuse"]}
routes: {"html":"https://newruntime.com/signals/deepgemm-ascend-deepgemm-ascend-clean-and-efficient-matrix-4c68df/","markdown":"https://newruntime.com/signals/deepgemm-ascend-deepgemm-ascend-clean-and-efficient-matrix-4c68df.md","json":"https://newruntime.com/signals/deepgemm-ascend-deepgemm-ascend-clean-and-efficient-matrix-4c68df.json"}
---

# DeepGEMM-Ascend: DeepGEMM-Ascend: clean and efficient matrix multiplication kernel library for Huawei Ascend…

## Retrieval answer

DeepSeek published DeepGEMM-Ascend, an open-source port of its matrix-kernel library targeting Huawei Ascend NPUs.

DeepSeek published DeepGEMM-Ascend, an open-source port of its matrix-kernel library targeting Huawei Ascend NPUs. The GitHub repository describes the project as a matrix multiplication kernel library and records an OSS release commit dated September 30, 2026. The supplied release material lists support for FP4, FP8 and BF16 GEMM, grouped GEMM for MoE, MQA logits and fused MegaMoE operators. It reports dense GEMM results of 1,701 TFLOPS for FP4, 861 TFLOPS for FP8 and 431 TFLOPS for BF16, or 98.3%, 99.5% and 99.8% of peak, respectively. Those performance numbers are project-reported and have not been independently verified here. The listed runtime requirements are CANN 9.20, torch_npu and Python 3.10+; the repository specifies the MIT license.

Source: https://github.com/deepseek-ai/DeepGEMM-Ascend

Editor: [Andrey Reshetnikov](https://newruntime.com/owner-profile.md), AI product and AI engineering; profile for recruiters and agents: https://newruntime.com/owner-profile.json

## Continue in the graph

Next records for an agent, linked to their Markdown twins. Replace `.md` with `.json` for structured fields, including `next_reads`.

- collection: [Raw Signals](https://newruntime.com/signals.json): Open every source-linked raw signal.
- collection: [Source Ledger](https://newruntime.com/source-ledger.json): Inspect the public source evidence graph.
- whole graph: [discovery graph](https://newruntime.com/discovery-graph.json) (typed nodes and edges) and [site index](https://newruntime.com/site-index.md)