mirror of https://github.com/opencv/opencv.git
Merge pull request #13070 from elatkin:el/gapi_perf_sobel
GAPI (fluid): optimization of Sobel 3x3 (#13070) * GAPI: performance test for Sobel * GAPI: performance test for Sobel w/FP32 input * GAPI: Sobel speedup: 2.5x (U8) up to 10x (float) * GAPI: Sobel 3x3 to support U8 into S16 * GAPI (fluid): Sobel 3x3 speedup: 10% (uchar), 1.5x (float) * GAPI (fluid): Sobel 3x3 speedup: +10x (uchar), but -20% (float) * GAPI (fluid): Sobel 3x3 speedup: +10% (float) * GAPI (fluid): Sobel 3x3 speedup: +15% (float), +10% (uchar) * GAPI (fluid): Sobel 3x3: address GCC warnings * GAPI (fluid): Sobel 3x3: separate *.cpp file w/SIMD code * GAPI (fluid): Sobel 3x3: fixed AVX2 code, AVX2 speedup 20-50% (uchar), 10-20% (float) * GAPI (fluid): Sobel 3x3: fix CV_SIMD code for AVX2 * GAPI (fluid): Sobel 3x3: refactorpull/13133/head^2
Evgeny Latkin
6 years ago
committed by
Alexander Alekhin
parent
a456b968cf
commit
4e40e5bb88
10 changed files with 420 additions and 51 deletions
@ -0,0 +1,38 @@ |
||||
// This file is part of OpenCV project.
|
||||
// It is subject to the license terms in the LICENSE file found in the top-level directory
|
||||
// of this distribution and at http://opencv.org/license.html.
|
||||
//
|
||||
// Copyright (C) 2018 Intel Corporation
|
||||
|
||||
|
||||
#include "../perf_precomp.hpp" |
||||
#include "../common/gapi_imgproc_perf_tests.hpp" |
||||
#include "../../src/backends/fluid/gfluidimgproc.hpp" |
||||
|
||||
|
||||
#define IMGPROC_FLUID cv::gapi::imgproc::fluid::kernels() |
||||
|
||||
namespace opencv_test |
||||
{ |
||||
|
||||
INSTANTIATE_TEST_CASE_P(SobelPerfTestFluid, SobelPerfTest, |
||||
Combine(Values(AbsExact().to_compare_f()), |
||||
Values(CV_8UC1, CV_8UC3, CV_16UC1, CV_16SC1), // add CV_32FC1 when ready
|
||||
Values(3), // add 5x5 once supported
|
||||
Values(szVGA, sz720p, sz1080p), |
||||
Values(-1, CV_16S, CV_32F), |
||||
Values(0, 1), |
||||
Values(1, 2), |
||||
Values(cv::compile_args(IMGPROC_FLUID)))); |
||||
|
||||
INSTANTIATE_TEST_CASE_P(SobelPerfTestFluid32F, SobelPerfTest, |
||||
Combine(Values(AbsToleranceSobel(1e-3).to_compare_f()), |
||||
Values(CV_32FC1), |
||||
Values(3), // add 5x5 once supported
|
||||
Values(szVGA, sz720p, sz1080p), |
||||
Values(CV_32F), |
||||
Values(0, 1), |
||||
Values(1, 2), |
||||
Values(cv::compile_args(IMGPROC_FLUID)))); |
||||
|
||||
} |
||||
@ -0,0 +1,270 @@ |
||||
// This file is part of OpenCV project.
|
||||
// It is subject to the license terms in the LICENSE file found in the top-level directory
|
||||
// of this distribution and at http://opencv.org/license.html.
|
||||
//
|
||||
// Copyright (C) 2018 Intel Corporation
|
||||
|
||||
#if !defined(GAPI_STANDALONE) |
||||
|
||||
#include "gfluidimgproc_func.hpp" |
||||
|
||||
#include "gfluidutils.hpp" |
||||
|
||||
#include <opencv2/core/hal/intrin.hpp> |
||||
|
||||
#include <cmath> |
||||
#include <cstdlib> |
||||
|
||||
#ifdef __GNUC__ |
||||
# pragma GCC diagnostic push |
||||
# pragma GCC diagnostic ignored "-Wstrict-overflow" |
||||
#endif |
||||
|
||||
namespace cv { |
||||
namespace gapi { |
||||
namespace fluid { |
||||
|
||||
//---------------------
|
||||
//
|
||||
// Fluid kernels: Sobel
|
||||
//
|
||||
//---------------------
|
||||
|
||||
// Sobel 3x3: vertical pass
|
||||
template<bool noscale, typename DST> |
||||
void run_sobel3x3_vert(DST out[], int length, const float ky[], |
||||
float scale, float delta, const int r[], float *buf[]) |
||||
{ |
||||
float ky0 = ky[0], |
||||
ky1 = ky[1], |
||||
ky2 = ky[2]; |
||||
|
||||
int r0 = r[0], |
||||
r1 = r[1], |
||||
r2 = r[2]; |
||||
|
||||
#if CV_SIMD |
||||
// for floating-point output,
|
||||
// manual vectoring may be not better than compiler's optimization
|
||||
#define EXPLICIT_SIMD_32F 0 // 1=vectorize 32f case explicitly, 0=don't
|
||||
#if EXPLICIT_SIMD_32F |
||||
if (std::is_same<DST, float>::value && length >= v_int16::nlanes) |
||||
{ |
||||
constexpr static int nlanes = v_float32::nlanes; |
||||
|
||||
for (int l=0; l < length; ) |
||||
{ |
||||
for (; l <= length - nlanes; l += nlanes) |
||||
{ |
||||
v_float32 sum = vx_load(&buf[r0][l]) * vx_setall_f32(ky0); |
||||
sum = v_fma(vx_load(&buf[r1][l]), vx_setall_f32(ky1), sum); |
||||
sum = v_fma(vx_load(&buf[r2][l]), vx_setall_f32(ky2), sum); |
||||
|
||||
if (!noscale) |
||||
{ |
||||
sum = v_fma(sum, vx_setall_f32(scale), vx_setall_f32(delta)); |
||||
} |
||||
|
||||
v_store(reinterpret_cast<float*>(&out[l]), sum); |
||||
} |
||||
|
||||
if (l < length) |
||||
{ |
||||
// tail: recalculate last pixels
|
||||
GAPI_DbgAssert(length >= nlanes); |
||||
l = length - nlanes; |
||||
} |
||||
} |
||||
|
||||
return; |
||||
} |
||||
#endif |
||||
|
||||
if ((std::is_same<DST, short>::value || std::is_same<DST, ushort>::value) |
||||
&& length >= v_int16::nlanes) |
||||
{ |
||||
constexpr static int nlanes = v_int16::nlanes; |
||||
|
||||
for (int l=0; l < length; ) |
||||
{ |
||||
for (; l <= length - nlanes; l += nlanes) |
||||
{ |
||||
v_float32 sum0 = vx_load(&buf[r0][l]) * vx_setall_f32(ky0); |
||||
sum0 = v_fma(vx_load(&buf[r1][l]), vx_setall_f32(ky1), sum0); |
||||
sum0 = v_fma(vx_load(&buf[r2][l]), vx_setall_f32(ky2), sum0); |
||||
|
||||
v_float32 sum1 = vx_load(&buf[r0][l + nlanes/2]) * vx_setall_f32(ky0); |
||||
sum1 = v_fma(vx_load(&buf[r1][l + nlanes/2]), vx_setall_f32(ky1), sum1); |
||||
sum1 = v_fma(vx_load(&buf[r2][l + nlanes/2]), vx_setall_f32(ky2), sum1); |
||||
|
||||
if (!noscale) |
||||
{ |
||||
sum0 = v_fma(sum0, vx_setall_f32(scale), vx_setall_f32(delta)); |
||||
sum1 = v_fma(sum1, vx_setall_f32(scale), vx_setall_f32(delta)); |
||||
} |
||||
|
||||
v_int32 isum0 = v_round(sum0), |
||||
isum1 = v_round(sum1); |
||||
|
||||
if (std::is_same<DST, short>::value) |
||||
{ |
||||
// signed short
|
||||
v_int16 res = v_pack(isum0, isum1); |
||||
v_store(reinterpret_cast<short*>(&out[l]), res); |
||||
} else |
||||
{ |
||||
// unsigned short
|
||||
v_uint16 res = v_pack_u(isum0, isum1); |
||||
v_store(reinterpret_cast<ushort*>(&out[l]), res); |
||||
} |
||||
} |
||||
|
||||
if (l < length) |
||||
{ |
||||
// tail: recalculate last pixels
|
||||
GAPI_DbgAssert(length >= nlanes); |
||||
l = length - nlanes; |
||||
} |
||||
} |
||||
|
||||
return; |
||||
} |
||||
|
||||
if (std::is_same<DST, uchar>::value && length >= v_uint8::nlanes) |
||||
{ |
||||
constexpr static int nlanes = v_uint8::nlanes; |
||||
|
||||
for (int l=0; l < length; ) |
||||
{ |
||||
for (; l <= length - nlanes; l += nlanes) |
||||
{ |
||||
v_float32 sum0 = vx_load(&buf[r0][l]) * vx_setall_f32(ky0); |
||||
sum0 = v_fma(vx_load(&buf[r1][l]), vx_setall_f32(ky1), sum0); |
||||
sum0 = v_fma(vx_load(&buf[r2][l]), vx_setall_f32(ky2), sum0); |
||||
|
||||
v_float32 sum1 = vx_load(&buf[r0][l + nlanes/4]) * vx_setall_f32(ky0); |
||||
sum1 = v_fma(vx_load(&buf[r1][l + nlanes/4]), vx_setall_f32(ky1), sum1); |
||||
sum1 = v_fma(vx_load(&buf[r2][l + nlanes/4]), vx_setall_f32(ky2), sum1); |
||||
|
||||
v_float32 sum2 = vx_load(&buf[r0][l + 2*nlanes/4]) * vx_setall_f32(ky0); |
||||
sum2 = v_fma(vx_load(&buf[r1][l + 2*nlanes/4]), vx_setall_f32(ky1), sum2); |
||||
sum2 = v_fma(vx_load(&buf[r2][l + 2*nlanes/4]), vx_setall_f32(ky2), sum2); |
||||
|
||||
v_float32 sum3 = vx_load(&buf[r0][l + 3*nlanes/4]) * vx_setall_f32(ky0); |
||||
sum3 = v_fma(vx_load(&buf[r1][l + 3*nlanes/4]), vx_setall_f32(ky1), sum3); |
||||
sum3 = v_fma(vx_load(&buf[r2][l + 3*nlanes/4]), vx_setall_f32(ky2), sum3); |
||||
|
||||
if (!noscale) |
||||
{ |
||||
sum0 = v_fma(sum0, vx_setall_f32(scale), vx_setall_f32(delta)); |
||||
sum1 = v_fma(sum1, vx_setall_f32(scale), vx_setall_f32(delta)); |
||||
sum2 = v_fma(sum2, vx_setall_f32(scale), vx_setall_f32(delta)); |
||||
sum3 = v_fma(sum3, vx_setall_f32(scale), vx_setall_f32(delta)); |
||||
} |
||||
|
||||
v_int32 isum0 = v_round(sum0), |
||||
isum1 = v_round(sum1), |
||||
isum2 = v_round(sum2), |
||||
isum3 = v_round(sum3); |
||||
|
||||
v_int16 ires0 = v_pack(isum0, isum1), |
||||
ires1 = v_pack(isum2, isum3); |
||||
|
||||
v_uint8 res = v_pack_u(ires0, ires1); |
||||
v_store(reinterpret_cast<uchar*>(&out[l]), res); |
||||
} |
||||
|
||||
if (l < length) |
||||
{ |
||||
// tail: recalculate last pixels
|
||||
GAPI_DbgAssert(length >= nlanes); |
||||
l = length - nlanes; |
||||
} |
||||
} |
||||
|
||||
return; |
||||
} |
||||
#endif |
||||
|
||||
// reference code
|
||||
for (int l=0; l < length; l++) |
||||
{ |
||||
float sum = buf[r0][l]*ky0 + buf[r1][l]*ky1 + buf[r2][l]*ky2; |
||||
|
||||
if (!noscale) |
||||
{ |
||||
sum = sum*scale + delta; |
||||
} |
||||
|
||||
out[l] = saturate<DST>(sum, rintf); |
||||
} |
||||
} |
||||
|
||||
template<typename DST, typename SRC> |
||||
void run_sobel_impl(DST out[], const SRC *in[], int width, int chan, |
||||
const float kx[], const float ky[], int border, |
||||
float scale, float delta, float *buf[], |
||||
int y, int y0) |
||||
{ |
||||
int r[3]; |
||||
r[0] = (y - y0) % 3; // buf[r[0]]: previous
|
||||
r[1] = (y - y0 + 1) % 3; // this
|
||||
r[2] = (y - y0 + 2) % 3; // next row
|
||||
|
||||
int length = width * chan; |
||||
|
||||
// horizontal pass
|
||||
|
||||
// full horizontal pass is needed only if very 1st row in ROI;
|
||||
// for 2nd and further rows, it is enough to convolve only the
|
||||
// "next" row - as we can reuse buffers from previous calls to
|
||||
// this kernel (note that Fluid processes rows consequently)
|
||||
int k0 = (y == y0)? 0: 2; |
||||
|
||||
for (int k = k0; k < 3; k++) |
||||
{ |
||||
// previous, this , next pixel
|
||||
const SRC *s[3] = {in[k] - border*chan , in[k], in[k] + border*chan}; |
||||
|
||||
// rely on compiler vectoring
|
||||
for (int l=0; l < length; l++) |
||||
{ |
||||
buf[r[k]][l] = s[0][l]*kx[0] + s[1][l]*kx[1] + s[2][l]*kx[2]; |
||||
} |
||||
} |
||||
|
||||
// vertical pass
|
||||
if (scale == 1 && delta == 0) |
||||
{ |
||||
constexpr static bool noscale = true; // omit scaling
|
||||
run_sobel3x3_vert<noscale, DST>(out, length, ky, scale, delta, r, buf); |
||||
} else |
||||
{ |
||||
constexpr static bool noscale = false; // do scaling
|
||||
run_sobel3x3_vert<noscale, DST>(out, length, ky, scale, delta, r, buf); |
||||
} |
||||
} |
||||
|
||||
#define INSTANTIATE(DST, SRC) \ |
||||
template void run_sobel_impl(DST out[], const SRC *in[], int width, int chan, \
|
||||
const float kx[], const float ky[], int border, \
|
||||
float scale, float delta, float *buf[], \
|
||||
int y, int y0); |
||||
|
||||
INSTANTIATE(uchar , uchar ) |
||||
INSTANTIATE(ushort, ushort) |
||||
INSTANTIATE( short, uchar ) |
||||
INSTANTIATE( short, ushort) |
||||
INSTANTIATE( short, short) |
||||
INSTANTIATE( float, uchar ) |
||||
INSTANTIATE( float, ushort) |
||||
INSTANTIATE( float, short) |
||||
INSTANTIATE( float, float) |
||||
|
||||
#undef INSTANTIATE |
||||
|
||||
} // namespace fliud
|
||||
} // namespace gapi
|
||||
} // namespace cv
|
||||
|
||||
#endif // !defined(GAPI_STANDALONE)
|
||||
@ -0,0 +1,31 @@ |
||||
// This file is part of OpenCV project.
|
||||
// It is subject to the license terms in the LICENSE file found in the top-level directory
|
||||
// of this distribution and at http://opencv.org/license.html.
|
||||
//
|
||||
// Copyright (C) 2018 Intel Corporation
|
||||
|
||||
#pragma once |
||||
|
||||
#if !defined(GAPI_STANDALONE) |
||||
|
||||
namespace cv { |
||||
namespace gapi { |
||||
namespace fluid { |
||||
|
||||
//---------------------
|
||||
//
|
||||
// Fluid kernels: Sobel
|
||||
//
|
||||
//---------------------
|
||||
|
||||
template<typename DST, typename SRC> |
||||
void run_sobel_impl(DST out[], const SRC *in[], int width, int chan, |
||||
const float kx[], const float ky[], int border, |
||||
float scale, float delta, float *buf[], |
||||
int y, int y0); |
||||
|
||||
} // namespace fluid
|
||||
} // namespace gapi
|
||||
} // namespace cv
|
||||
|
||||
#endif // !defined(GAPI_STANDALONE)
|
||||
Loading…
Reference in new issue