Skip to content

[CPU][RV64] implement vectorized im2col+GEMM convolution with fallback #35949

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Open
strelkovkm wants to merge 19 commits into
base: master
Choose a base branch
from
+23 −7
Open

[CPU][RV64] implement vectorized im2col+GEMM convolution with fallback #35949

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
19 commits
Select commit Hold shift + click to select a range
cc5809d
Edit: gitignore
strelkovkm Feb 18, 2026
fe1ad6a
[CPU] rv64: Add rvv gemm conv
strelkovkm Mar 13, 2026
bc6a5f5
[CPU][RV64] Edit: convolution_implementations.cpp - add gemm_conv
strelkovkm Mar 20, 2026
f9edc4f
[CPU][RV64] Bugfix: jit_conv_1x1
strelkovkm Mar 23, 2026
2387986
[CPU][RV64] Return ref + Add 1x1 conv
strelkovkm Apr 22, 2026
0aaa56c
[CPU][RV64] clean whitespace changes & rename impl at convolution_imp…
strelkovkm Apr 26, 2026
3d4c8ef
[CPU][RV64] Return ref and edit choose impl
strelkovkm Apr 28, 2026
db6022a
[CPU][RV64] Remove gemm ref switch
strelkovkm Apr 28, 2026
1039331
[CPU][RV64] Clean up dnnl
strelkovkm Apr 28, 2026
dc24375
[CPU][RV64] Clean up conv_impl
strelkovkm Apr 28, 2026
beb7797
[CPU][RV64] Clean up conv_impl
strelkovkm Apr 28, 2026
c66f745
[CPU][RV64] Clean up conv_impl
strelkovkm Apr 28, 2026
8c47a89
[CPU][RV64] Clean up conv_impl
strelkovkm Apr 28, 2026
d9c0a6c
[CPU][RV64] Clean up conv_impl
strelkovkm Apr 28, 2026
5738884
[CPU][RV64] Clean up conv_impl
strelkovkm Apr 28, 2026
35fd574
Merge branch 'master' into dev_conv_rv64
strelkovkm May 9, 2026
19e2fc7
Merge Master
strelkovkm May 9, 2026
c6bf4af
Merge master, keeping local dev_conv_rv64 version of oneDNN
strelkovkm May 14, 2026
66678bb
[CPU][RV64] clear /.gitignore
strelkovkm May 26, 2026
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
2 changes: 1 addition & 1 deletion .gitignore
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -68,4 +68,4 @@ coverage
/tools/commit_slider/*.json
/tools/commit_slider/slider_cache/*
/src/plugins/intel_cpu/thirdparty/ComputeLibrary/build/arm_compute_version.embed
.github/GITHUB_OUTPUT
.github/GITHUB_OUTPUT
2 changes: 1 addition & 1 deletion src/bindings/python/thirdparty/pybind11
Open in desktop
Submodule pybind11 updated 142 files
16 changes: 16 additions & 0 deletions src/plugins/intel_cpu/src/nodes/executors/convolution_implementations.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -276,6 +276,22 @@ const std::vector<ExecutorImplementation<ConvAttrs>>& getImplementations() {
AcceptsAnyShape<ConvAttrs>,
CreateDnnlDefault<DnnlConvolutionPrimitive, ConvAttrs>{}
)
OV_CPU_INSTANCE_RISCV64(
"convolution_dnnl_ncsp_ncsp", ExecutorType::Dnnl, OperationType::Convolution,
// supports
[](const ConvConfig& config, const MemoryFormatFilter& memoryFormatFilter) -> bool {
VERIFY(!isQuantized(config), UNSUPPORTED_SRC_PRECISIONS);
Comment on lines +279 to +283
VERIFY(!hasPostOp<DepthwiseConvolutionPostOp>(config.attrs.postOps), UNSUPPORTED_POST_OPS);
return MatchesMemoryFormatFilter(config.descs,
LayoutConfig{LayoutType::ncsp, LayoutType::ncsp, LayoutType::ncsp, LayoutType::ncsp},
memoryFormatFilter,
dnnlConvolutionMappingNotation);
},
// createOptimalConfig
CreateOptimalConfigDefault{{LayoutType::ncsp, LayoutType::ncsp, LayoutType::ncsp, LayoutType::ncsp}},
AcceptsAnyShape<ConvAttrs>,
CreateDnnlDefault<DnnlConvolutionPrimitive, ConvAttrs>{}
Comment on lines +291 to +293
)
OV_CPU_INSTANCE_RISCV64(
"convolution_dnnl_ref_ncsp", ExecutorType::Dnnl, OperationType::Convolution,
// supports
Expand Down
2 changes: 1 addition & 1 deletion src/plugins/intel_cpu/thirdparty/ComputeLibrary
Open in desktop
Submodule ComputeLibrary updated 206 files
2 changes: 1 addition & 1 deletion src/plugins/intel_cpu/thirdparty/onednn
Open in desktop
Submodule onednn updated 10 files
+1 −1 src/cpu/CMakeLists.txt
+7 −0 src/cpu/cpu_convolution_list.cpp
+585 −0 src/cpu/rv64/rvv_gemm_convolution.cpp
+150 −0 src/cpu/rv64/rvv_gemm_convolution.hpp
+2,197 −0 src/cpu/rv64/rvv_gemm_convolution_utils.cpp
+142 −0 src/cpu/rv64/rvv_gemm_convolution_utils.hpp
+6 −7 src/cpu/x64/jit_brgemm_1x1_conv.cpp
+23 −24 src/cpu/x64/jit_brgemm_conv.cpp
+14 −36 src/cpu/x64/jit_brgemm_conv_utils.cpp
+2 −3 src/cpu/x64/jit_brgemm_conv_utils.hpp
2 changes: 1 addition & 1 deletion src/plugins/intel_gpu/thirdparty/onednn_gpu
Open in desktop
Submodule onednn_gpu updated from 20db47 to 7e6e4d
2 changes: 1 addition & 1 deletion src/plugins/intel_npu/thirdparty/level-zero-ext
Open in desktop
Submodule level-zero-ext updated 1 files
+1 −5 ze_graph_ext.h
2 changes: 1 addition & 1 deletion thirdparty/gtest/gtest
Open in desktop
Copy link
Copy Markdown
Contributor

@aobolensk aobolensk May 16, 2026

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Please, get rid of all dependency updates, except onednn (CPU only)

Submodule gtest updated 295 files
Loading