uxlfoundation/oneDNN v3.12.1

on GitHub

This is a patch release containing the following changes to v3.12:

• Enabled SYCL Graph record/replay mode support in Graph API on Intel GPUs (f014724, 3e19991, 982650a)
• Fixed a performance regression in matmul with 4D shapes and N == 1 or M == 1 on x64 CPUs (de9498f, 540d53d)
• Fixed correctness issue in matmul primitive with ReLU post-op on RV64 CPUs (772ca13)
• Fixed a segfault in s8/u8 depthwise convolution on x64 processors with Intel AVX10.2 instruction set support (7f413f9, dac7399)
• Fixed a correctness issue in s8s8 convolution on x64 processors with Intel AVX-512 and Intel DL Boost instructions support (0dc2ca8)
• Fixed an issue with incorrect memory use estimation of layer normalization, group normalization, and batch normalization primitives in benchdnn (881e9b6)
• Fixed an assertion in benchdnn --graph driver (79b2593)