Skip to content

Add conv im2col int8 #15

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
Xinyu302 wants to merge 11 commits into
base: add-basic-conv
Choose a base branch
from
Open

Add conv im2col int8 #15

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
11 commits
Select commit Hold shift + click to select a range
d87175d
Merge pull request #14 from Xinyu302/add-basic-conv
Xinyu302 Feb 18, 2024
2795f0e
fix net.cpp layer pack
Xinyu302 Feb 18, 2024
1d945be
Merge branch 'rvv-int8' into add-depthwise-conv
Xinyu302 Feb 18, 2024
14e53f6
fix segfault bug
Xinyu302 Feb 18, 2024
7013427
WIP: buggy int8 packn
Xinyu302 Feb 20, 2024
d049bc3
WIP: maybe fixed
Xinyu302 Feb 20, 2024
03156e8
apply code-format changes
fty1777 Feb 20, 2024
280ce89
Merge branch 'add-depthwise-conv' into version-2-21
Xinyu302 Feb 21, 2024
2c0d165
Merge branch 'add-depthwise-conv' into add-conv-im2col-int8
Xinyu302 Feb 21, 2024
ddf2329
fix depthwise conv bug
Xinyu302 Feb 21, 2024
64009eb
Merge branches 'add-conv-im2col-int8' and 'add-conv-im2col-int8' of h…
Xinyu302 Feb 21, 2024
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
68 changes: 68 additions & 0 deletions src/layer/riscv/convolution_1x1_packn_int8.h
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -0,0 +1,68 @@
// Tencent is pleased to support the open source community by making ncnn available.
//
// Copyright (C) 2021 THL A29 Limited, a Tencent company. All rights reserved.
//
// Licensed under the BSD 3-Clause License (the "License"); you may not use this file except
// in compliance with the License. You may obtain a copy of the License at
//
// https://opensource.org/licenses/BSD-3-Clause
//
// Unless required by applicable law or agreed to in writing, software distributed
// under the License is distributed on an "AS IS" BASIS, WITHOUT WARRANTIES OR
// CONDITIONS OF ANY KIND, either express or implied. See the License for the
// specific language governing permissions and limitations under the License.

static void conv1x1s1_sgemm_packn_int8_rvv(const Mat& bottom_blob, Mat& top_blob, const Mat& kernel, const Option& opt)
{
int w = bottom_blob.w;
int h = bottom_blob.h;
const int size = w * h;

Mat bottom_im2col = bottom_blob;
bottom_im2col.w = size;
bottom_im2col.h = 1;

im2col_sgemm_packn_int8_rvv(bottom_im2col, top_blob, kernel, opt);
}

static void conv1x1s2_sgemm_packn_int8_rvv(const Mat& bottom_blob, Mat& top_blob, const Mat& kernel, const Option& opt)
{
const int packn = csrr_vlenb() / 2;
const size_t vl = vsetvl_e8m1(packn);

int w = bottom_blob.w;
int channels = bottom_blob.c;
size_t elemsize = bottom_blob.elemsize;
int elempack = bottom_blob.elempack;

int outw = top_blob.w;
int outh = top_blob.h;

const int tailstep = (w - 2 * outw + w) * packn;

Mat bottom_blob_shrinked;
bottom_blob_shrinked.create(outw, outh, channels, elemsize, elempack, opt.workspace_allocator);

#pragma omp parallel for num_threads(opt.num_threads)
for (int p = 0; p < channels; p++)
{
const int8_t* r0 = bottom_blob.channel(p);
int8_t* outptr = bottom_blob_shrinked.channel(p);

for (int i = 0; i < outh; i++)
{
for (int j = 0; j < outw; j++)
{
vint8m1_t _val = vle8_v_i8m1(r0, vl);
vse8_v_i8m1(outptr, _val, vl);

r0 += packn * 2;
outptr += packn;
}

r0 += tailstep;
}
}

conv1x1s1_sgemm_packn_int8_rvv(bottom_blob_shrinked, top_blob, kernel, opt);
}
Loading