uxlfoundation / oneDNN
Open-source library of optimized deep learning operations (matmul, convolution, attention) for CPUs (x64, AArch64, RISC-V) and Intel GPUs. Used by PyTorch, TensorFlow, OpenVINO, and ONNX Runtime.
Why this job: works in the areas we cover, but does not say which job it does
What it does
- Single API, many devices. The same code targets Intel 64/AMD64, AArch64,
- Hardware specialization. oneDNN detects hardware features at runtime
- Low-precision inference. Supports f32, bf16, f16, f8, f4,
- Operation fusion. Post-ops fuse activations, residual connections, and
- Interoperability. Supports OpenMP, TBB, and custom threadpools on CPUs,
Get started
conda install -c conda-forge onednnYou need: Intel 64/AMD64; Arm 64-bit Architecture (AArch64); RISC-V 64-bit (RV64)
- Maintenance
- Active
- Last commit today
- Stars
- 4,092
- +2 this week
- Licence
- Apache-2.0
- Allows business use
- Downloads for
- None listed
- No release files
- Latest version
- v3.14.2
- Released 8 days ago
- Built with
- C++
- 95% of the code
Is it looked after?
- Contributors
- 16
- Last 90 days
- Issues
- 97% closed
- 40 still open
- Pull requests
- 4,190 merged
- 139 waiting
- Releases
- 214
- 26 in the last year
More: languages, topics and project record
Languages
- C++ 95%
- C 4%
- CMake 1%
- Python <1%
Topics
- onednn
- oneapi
- performance
- cpp
- x64
- aarch64
- amx
- sycl
- ace
- convolution
- inference
- intel-gpu
Project record
- JobOther AI tools
- StatusListed
- Commits24,545
- Forks1,218
- Watching157
- Stars in 28 days+32
- CreatedMay 9, 2016
- First seen by usNov 7, 2026
- CheckedNov 12, 2026