Download Latest Version
v3.10.2 source code.tar.gz (13.5 MB)
Get an email when there's a new version of oneDNN
| Name | Modified | Size | Downloads / Week |
|---|---|---|---|
| Parent folder | |||
| README.md | 2025-12-02 | 1.0 kB | |
| v3.10.2 source code.tar.gz | 2025-12-02 | 13.5 MB | |
| v3.10.2 source code.zip | 2025-12-02 | 16.8 MB | |
| Totals: 3 Items | 30.3 MB | 1 | |
This is a patch release containing the following changes to v3.10.1:
* Fixed a memory leak in Graph API related to host scalars use (044124555fbf17c37196e4a565d61c10d8640bfc)
* Fixed f16 matmul performance regression with int4 weights on Intel Arc graphics for Intel Core Ultra processors (Series 3) (789711cc8dc1b1e163ce74f289849150464395e1, a1602472f5ee3d62dc0b3394604b98b4290ceeb7)
* Fixed bf16 matmul performance regression on Intel Xeon processors with Intel AMX instruction set support (c29ec26019da8bcd80b4e28ba33e5d47d0439333)
* Changed register allocation in BRGEMM kernel to avoid register conflicts and improve code safety (95d651be16384dea97e140ca579a9c79e5e3862d)
* Fixed a crash related to incorrect caching of int8 convolution primitive on Intel GPUs (28ccca41846882ef741d53c8adf20ed25406c0df, 0bc8060b68a27d5f9ee00b160a83ff9b7010a154)
* Fixed a bug preventing correct detection of Intel AVX 10.2 instruction set on Intel Xeon processors (568171cbe245a06ae68bcf676476c12805d2468e)
