oneDNN Files

This is a patch release containing the following changes to v3.13.2:

• Improved performance of Scaled Dot Product Attention (SDPA) training forward subgraph with Graph API on x64 CPUs (e971e88e3e58bbf6d033f2b7796cb4d3aa354715, b5206d31a903e7b82420095f05619cbe5d9139fe, bf8ffaa2d3105a8c473cf52f498b70716904b386)
• Fixed build errors with -Wunused-template diagnostic promoted to an error (75c3dfa0c0baa1e6f8ce58236b9d13783698b5d5, 6e7312c30ffaa5f8b2d8888fb8e5a08e6235e0ed)
• Improved performance of SDPA backpropagation on Intel GPUs (fc9fed9244f401e05ef772aef105a4cdc7848679, 695317ec226c61b4520f08f311545e69ef1ae7a7)
• Introduced dnnl::verbose_profiling_enabled function to allow applications to check whether verbose profiling mode requires queue profiling to be enabled (e2a772f8837a6275eee001cebb45d2dc05985112, d005652756249b4598e8e246c99c8337ddbe83cc, 1716509706ed075cde61781d00aaf3784ad5a840)
• Fixed crash in convolution with large padding on CPUs with Intel AVX-512 and Intel AVX2 instruction set support (3c1407aafe5669269ec3463306f8e805e34cd310)
• Fixed crash and hang in verbose mode with Graph API and ONEDNN_CPU_RUNTIME=THREADPOOL when using an asynchronous threadpool (a35fd4f0abfedeef3df6e7a2cebc0ced2210a105, 25b6be636b0ed088e3af4a764751c1d3f7b146b1, a1c07f21d577e79b5df50faa262653cd0b8e6f61, b7dc0d5b49b68ed7d65c82822eacf341fee07eeb)
• Fixed sporadic correctness issue in SDPA subgraph on Intel GPUs (b38b47c9af53e2ad744241613728fa6cc545ec06)
• Fixed correctness issue in RNN primitive backpropagation with GRU cell type and dhc == 1 on Intel GPUs (0e486f6520521028cbc4aca5e81d37c17de292a0)
• Fixed a crash in matmul with transposed tensor B and non-trivial strides on x64 CPUs (c0abb1f69e9adba79ca1497d2cc1f8c8553dadda)
• [experimental] Introduced support for verbose profiling for SYCL runtime based on sycl_ext_oneapi_profiling_tag extension (6a45b07b05c435e7a054925d790b54f036235c61, 303e8174e5f856501b0dd1851f0974eea996aa48, e26fd5ab54663d7f2eb677a1649f920261c70c93, 3fc4f4ef6e1f647580ccaa6601ae63608e834751, 8f8e9ecae4388a59a840b2b91c13773167e693d3, bdf5118740660b1b490ece6b90a1f98e7bf286bf, bdf5118740660b1b490ece6b90a1f98e7bf286bf, 4f1cd922c7ca4e873ad88c5cd43615813a5bbbfb, 13d58e9d4df6c4e762a45d0effe80ed657eba564)
• Fixed crash in matmul with binary post-ops and bf16 or f16 broadcasted tensors on x64 CPUs with Intel AVX-512 and Intel DL Boost support (933b30af0964b0ceee8f0124221764343272ee10, b29c933ffd5d090187cd4392977824a509dfc637)
• Fixed correctness issues in f32 3D matmul with fp16 or bf16 weights on x64 CPUs (68f6a269ee53715b2036c61669dbcdc6a9c63bb6, 171872ed9eb22d5c536c1126a26534ae05811262)
• Fixed f32 SDPA subgraph performance regression on Intel GPUs (1a38c2dc5c15efc7d4e77f2d1c05e17b93354a27)
• Fixed performance regression in f32 convolutions with 1x1 kernel on AArch64 CPUs with SVE support (bb929172408c4d944f4172614e5d78cca2725605)
• Fixed performance regression in batch normalization on AArch64 CPUs with SVE support (a72c4a7543e3a1aeead4aeae7c963010641a6489)
• Changed Clang compiler flag from -fstack-protector-all to -fstack-protector-strong for builds on AArch64 CPUs (27fc2c124a299b98d4b890533d2740ed9316248f)
• Fixed correctness issue in f32 matmul with binary add post-op preceding sum post-op on Intel GPUs (f1c9e4367a892ab386edaa2ec11f82b57c1c1e4a)