onednn
https://github.com/oneapi-src/onednn
C++
oneAPI Deep Neural Network Library (oneDNN)
Triage Issues!
When you volunteer to triage issues, you'll receive an email each day with a link to an open issue that needs help in this project. You'll also receive instructions on how to triage issues.
Triage Docs!
Receive a documented method or class from your favorite GitHub repos in your inbox every day. If you're really pro, receive undocumented methods or classes and supercharge your commit history.
C++ not yet supported2 Subscribers
Add a CodeTriage badge to onednn
Help out
- Issues
- rfcs: rv64: add rvcl unified computation library integration
- aarch64: use register manager for injectors
- cpu: x64: matmul: restore MXFP4 weights decompression for f32
- cpu: aarch64: remove inappropriate uses of get_max_cpu_isa()
- [WIP]: x64: matmul: performance improvements on int4 grouped quantization
- Arc Pro B65: FP16 MatMul produces Infinity via OpenVINO GPU; disabling oneDNN restores finite outputs
- CPU x64: Guidance on creating a matmul primitive at inside primitive execute for an implementation that owns an opaque prepacked weights format (zen64)
- cpu: gemm: include only CBLAS from Accelerate
- cpu: x64: enable preconvert fp8 to f16 for conv/deconv on NVL
- [AArch64][Perf] Regressions for BF16 MatMuls with kleidi ops and newer versions of ACL
- Docs
- C++ not yet supported