← Discover
GitHub · deepseek-ai/DeepGEMM↑ +363 today

Fast matrix computation for artificial intelligence

DeepGEMM, developed by DeepSeek, is an open-source basic linear algebra subprograms (BLAS) library that accelerates matrix multiplication operations on graphics processing units (GPUs). The software provides optimized kernels for artificial intelligence models requiring high-performance computing.

Updates

  • October 6, 2026: Stars 8,522 → 8,528, latest release v2.1.1.post3 (October 15, 2025).

What you get

  • It shortens the execution times of large language models by accelerating matrix multiplications.
  • It automatically compiles kernels at runtime without waiting for CUDA compilation during installation.
  • It reduces graphics card communication losses by combining different expert models in a single operation.

Installation

Cloning the repository and the development environment h
# Submodule must be cloned
git clone --recursive git@github.com:deepseek-ai/DeepGEMM.git
cd DeepGEMM

# Link some essential includes and build the C++ extension
cat develop.sh
./develop.sh
Installing the library
cat install.sh
./install.sh

If you don't write code

Paste this into your AI agent (Claude Code · Codex · Antigravity)

I want to install the DeepGEMM library on my hardware with NVIDIA SM90 or SM100 architecture. First, run the commands '# Submodule must be cloned\ngit clone --recursive git@github.com:deepseek-ai/DeepGEMM.git\ncd DeepGEMM\n\n# Link some essential includes and build the C++ extension\ncat develop.sh\n./develop.sh' to clone the repository with its submodules and prepare the development environment. Then, execute the command 'cat install.sh\n./install.sh' to complete the installation, and make the library ready for use in the Python environment with the 'import deep_gemm' command.

Repository: https://github.com/deepseek-ai/DeepGEMM. Follow the official instructions in the README of this repository for setup and execution.

Before starting, check which OS we are on and if the necessary tools (e.g., Python, Node.js, Git, Docker) are installed. Install any missing ones step-by-step according to this system.

Create a new, empty folder to work in. Do not touch any files outside this folder and do not delete anything. Tell me what you are going to do in one sentence before running each command.

Tell me what I should see on the screen at the end of each step. If an error occurs, I will paste the error message to you exactly as it is, and we will fix it together.

Related dictionary terms

BLASKernelsCloneGPUArtificial Intelligence
Who it is forIt is for developers and researchers who want to run large artificial intelligence models with the highest performance on modern NVIDIA graphics processing units.
LicenseMIT

Links

TreScout did not build this tool · we found it in GitHub trends and wrote it up. This page describes the repository as of 2026-10-06: The star count and our text belong to that day, the repository may have changed since. Check the repository link for the current state. The introductory text was prepared with AI from the repository's own documentation, and this page was machine-translated from the Turkish original · the Turkish version prevails.