Skip to content

Commit

Permalink
[GPU] Update onednn to v2.6-pc (openvinotoolkit#9479)
Browse files Browse the repository at this point in the history
  • Loading branch information
@isanghao
isanghao authored Dec 30, 2021
1 parent c23025d commit 5beb5dc
Showing 1 changed file with 1 addition and 1 deletion.
2 changes: 1 addition & 1 deletion src/plugins/intel_gpu/thirdparty/onednn_gpu
Submodule onednn_gpu updated 98 files
+8 −4 doc/build/build.md
+24 −23 doc/performance_considerations/profilers.md
+5 −4 doc/primitives/batch_normalization.md
+16 −12 doc/primitives/binary.md
+5 −4 doc/primitives/concat.md
+3 −2 doc/primitives/convolution.md
+5 −4 doc/primitives/eltwise.md
+4 −3 doc/primitives/inner_product.md
+4 −3 doc/primitives/layer_normalization.md
+5 −3 doc/primitives/logsoftmax.md
+3 −2 doc/primitives/lrn.md
+19 −8 doc/primitives/matmul.md
+4 −3 doc/primitives/pooling.md
+6 −3 doc/primitives/prelu.md
+7 −5 doc/primitives/reduction.md
+4 −3 doc/primitives/reorder.md
+7 −5 doc/primitives/resampling.md
+7 −6 doc/primitives/rnn.md
+5 −6 doc/primitives/shuffle.md
+3 −2 doc/primitives/softmax.md
+5 −3 doc/primitives/sum.md
+2 −0 include/oneapi/dnnl/dnnl.hpp
+1 −0 include/oneapi/dnnl/dnnl_types.h
+6 −10 src/common/binary.cpp
+1 −0 src/common/c_types_map.hpp
+1 −0 src/common/dnnl_debug_autogenerated.cpp
+1 −0 src/common/memory_desc_wrapper.cpp
+7 −3 src/common/utils.hpp
+14 −22 src/cpu/x64/brgemm/jit_brgemm_amx_uker.cpp
+31 −26 src/cpu/x64/gemm/f32/jit_avx512_common_gemm_f32.cpp
+2 −2 src/cpu/x64/gemm/f32/jit_avx512_common_gemm_f32.hpp
+35 −27 src/cpu/x64/gemm/f32/jit_avx_gemm_f32.cpp
+2 −2 src/cpu/x64/gemm/f32/jit_avx_gemm_f32.hpp
+96 −50 src/cpu/x64/gemm/gemm_driver.cpp
+14 −9 src/cpu/x64/injectors/jit_uni_eltwise_injector.cpp
+10 −6 src/cpu/x64/injectors/jit_uni_eltwise_injector.hpp
+2 −1 src/cpu/x64/injectors/jit_uni_postops_injector.cpp
+7 −6 src/cpu/x64/jit_uni_batch_normalization_s8.cpp
+5 −1 src/cpu/x64/jit_uni_tbb_batch_normalization.cpp
+46 −11 src/cpu/x64/matmul/brgemm_matmul.cpp
+14 −7 src/cpu/x64/matmul/brgemm_matmul_copy_utils.cpp
+66 −15 src/cpu/x64/matmul/brgemm_matmul_utils.cpp
+11 −1 src/cpu/x64/matmul/brgemm_matmul_utils.hpp
+14 −0 src/gpu/compute/device_info.cpp
+2 −0 src/gpu/compute/device_info.hpp
+19 −0 src/gpu/gpu_reorder_pd.hpp
+4 −0 src/gpu/jit/conv/bank_conflict_allocation.cpp
+267 −153 src/gpu/jit/conv/config.hpp
+250 −148 src/gpu/jit/conv/conv_kernel.hpp
+15 −14 src/gpu/jit/conv/fma_support.cpp
+1 −1 src/gpu/jit/conv/gemm_schedule.hpp
+18 −2 src/gpu/jit/conv/gen_convolution.cpp
+30 −32 src/gpu/jit/conv/ir.cpp
+46 −12 src/gpu/jit/conv/ir.hpp
+46 −0 src/gpu/jit/conv/ir_core.hpp
+22 −3 src/gpu/jit/conv/ir_simplify.cpp
+299 −23 src/gpu/jit/conv/kernel_builder.cpp
+6 −1 src/gpu/jit/conv/message_support.hpp
+6 −7 src/gpu/jit/conv/post_op_support.hpp
+21 −1 src/gpu/jit/conv/reg_buf.hpp
+14 −0 src/gpu/jit/conv/register_allocator.hpp
+22 −16 src/gpu/jit/conv/tensor.cpp
+18 −5 src/gpu/jit/conv/tensor.hpp
+13 −1 src/gpu/jit/gemm/emulation.hpp
+1 −0 src/gpu/jit/gemm/gemm_recipes.hpp
+5 −0 src/gpu/jit/gemm/gen_gemm.cpp
+6 −2 src/gpu/jit/gemm/gen_gemm.hpp
+10 −9 src/gpu/jit/gemm/gen_gemm_kernel.cpp
+571 −200 src/gpu/jit/gemm/gen_gemm_kernel_generator.cpp
+15 −3 src/gpu/jit/gemm/gen_gemm_kernel_generator.hpp
+3 −2 src/gpu/jit/gemm/xe_hp_systolic_gemm.hpp
+4 −2 src/gpu/nvidia/sycl_cuda_compat.hpp
+2 −1 src/gpu/ocl/cross_engine_reorder.cpp
+2 −4 src/gpu/ocl/cross_engine_reorder.hpp
+2 −2 src/gpu/ocl/custom_reorder.cl
+10 −7 src/gpu/ocl/custom_reorder.cpp
+5 −25 src/gpu/ocl/custom_reorder.hpp
+15 −13 src/gpu/ocl/gemm/gemm_with_post_ops.cl
+4 −2 src/gpu/ocl/gemm/gemm_with_post_ops.cpp
+12 −0 src/gpu/ocl/gemm/gen9_gemm.hpp
+5 −25 src/gpu/ocl/generic_reorder.hpp
+1 −5 src/gpu/ocl/ocl_eltwise.h
+1 −1 src/gpu/ocl/ref_matmul.hpp
+5 −25 src/gpu/ocl/ref_reorder.hpp
+15 −13 src/gpu/ocl/xe_lp_conv_dw_fwd_data_ow_block_x8s8x.cl
+13 −11 src/gpu/ocl/xe_lp_conv_nhwc_fwd_dw_ow_block_x8s8x.cl
+0 −1 src/gpu/primitive_conf.hpp
+8 −0 src/sycl/sycl_compat.cpp
+1 −1 tests/benchdnn/conv/cfg.cpp
+2 −1 tests/benchdnn/dnnl_common.cpp
+1 −0 tests/benchdnn/dnnl_debug_autogenerated.cpp
+4 −0 tests/benchdnn/inputs/concat/test_concat_gpu
+5 −8 tests/benchdnn/inputs/conv/test_conv_gpu
+14 −0 tests/benchdnn/inputs/matmul/test_matmul_multidims
+17 −2 tests/benchdnn/inputs/reorder/harness_reorder_compensation
+15 −10 tests/benchdnn/reorder/reorder.cpp
+2 −2 tests/gtests/gtest/gtest-matchers.h
+12 −0 tests/gtests/in/gemm_in.h

0 comments on commit 5beb5dc

Please sign in to comment.