mirror/opencv
1
0
Fork 0
mirror of https://github.com/opencv/opencv.git synced 2024-12-16 10:29:11 +08:00
Code Issues Actions Packages Projects Releases Wiki Activity
89d3f95a8e
opencv/modules/core/src/convert_scale.simd.hpp

365 lines
14 KiB
C++
Raw Normal View History

Split convert.cpp into smaller pieces
2018-02-06 20:02:51 +08:00
// 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
#include "precomp.hpp"
#include "convert.hpp"
core: dispatch convert_scale
2019-02-08 22:56:25 +08:00
namespace cv {
CV_CPU_OPTIMIZATION_NAMESPACE_BEGIN
BinaryFunc getCvtScaleAbsFunc(int depth);
BinaryFunc getConvertScaleFunc(int sdepth, int ddepth);
#ifndef CV_CPU_OPTIMIZATION_DECLARATIONS_ONLY
Split convert.cpp into smaller pieces
2018-02-06 20:02:51 +08:00
/****************************************************************************************\
* convertScale[Abs] *
\****************************************************************************************/
Merge pull request #12411 from vpisarev:wide_convert * rewrote Mat::convertTo() and convertScaleAbs() to wide universal intrinsics; added always-available and SIMD-optimized FP16<=>FP32 conversion * fixed compile warnings * fix some more compile errors * slightly relaxed accuracy threshold for int->float conversion (since we now do it using single-precision arithmetics, not double-precision) * fixed compile errors on iOS, Android and in the baseline C++ version (intrin_cpp.hpp) * trying to fix ARM-neon builds * trying to fix ARM-neon builds * trying to fix ARM-neon builds * trying to fix ARM-neon builds
2018-09-07 00:36:59 +08:00
template<typename _Ts, typename _Td> inline void
cvtabs_32f( const _Ts* src, size_t sstep, _Td* dst, size_t dstep,
Size size, float a, float b )
Split convert.cpp into smaller pieces
2018-02-06 20:02:51 +08:00
{
Merge pull request #12411 from vpisarev:wide_convert * rewrote Mat::convertTo() and convertScaleAbs() to wide universal intrinsics; added always-available and SIMD-optimized FP16<=>FP32 conversion * fixed compile warnings * fix some more compile errors * slightly relaxed accuracy threshold for int->float conversion (since we now do it using single-precision arithmetics, not double-precision) * fixed compile errors on iOS, Android and in the baseline C++ version (intrin_cpp.hpp) * trying to fix ARM-neon builds * trying to fix ARM-neon builds * trying to fix ARM-neon builds * trying to fix ARM-neon builds
2018-09-07 00:36:59 +08:00
#if CV_SIMD
v_float32 va = vx_setall_f32(a), vb = vx_setall_f32(b);
const int VECSZ = v_float32::nlanes*2;
Split convert.cpp into smaller pieces
2018-02-06 20:02:51 +08:00
#endif
Merge pull request #12411 from vpisarev:wide_convert * rewrote Mat::convertTo() and convertScaleAbs() to wide universal intrinsics; added always-available and SIMD-optimized FP16<=>FP32 conversion * fixed compile warnings * fix some more compile errors * slightly relaxed accuracy threshold for int->float conversion (since we now do it using single-precision arithmetics, not double-precision) * fixed compile errors on iOS, Android and in the baseline C++ version (intrin_cpp.hpp) * trying to fix ARM-neon builds * trying to fix ARM-neon builds * trying to fix ARM-neon builds * trying to fix ARM-neon builds
2018-09-07 00:36:59 +08:00
sstep /= sizeof(src[0]);
dstep /= sizeof(dst[0]);
Split convert.cpp into smaller pieces
2018-02-06 20:02:51 +08:00
Merge pull request #12411 from vpisarev:wide_convert * rewrote Mat::convertTo() and convertScaleAbs() to wide universal intrinsics; added always-available and SIMD-optimized FP16<=>FP32 conversion * fixed compile warnings * fix some more compile errors * slightly relaxed accuracy threshold for int->float conversion (since we now do it using single-precision arithmetics, not double-precision) * fixed compile errors on iOS, Android and in the baseline C++ version (intrin_cpp.hpp) * trying to fix ARM-neon builds * trying to fix ARM-neon builds * trying to fix ARM-neon builds * trying to fix ARM-neon builds
2018-09-07 00:36:59 +08:00
for( int i = 0; i < size.height; i++, src += sstep, dst += dstep )
Split convert.cpp into smaller pieces
2018-02-06 20:02:51 +08:00
{
Merge pull request #12411 from vpisarev:wide_convert * rewrote Mat::convertTo() and convertScaleAbs() to wide universal intrinsics; added always-available and SIMD-optimized FP16<=>FP32 conversion * fixed compile warnings * fix some more compile errors * slightly relaxed accuracy threshold for int->float conversion (since we now do it using single-precision arithmetics, not double-precision) * fixed compile errors on iOS, Android and in the baseline C++ version (intrin_cpp.hpp) * trying to fix ARM-neon builds * trying to fix ARM-neon builds * trying to fix ARM-neon builds * trying to fix ARM-neon builds
2018-09-07 00:36:59 +08:00
int j = 0;
#if CV_SIMD
for( ; j < size.width; j += VECSZ )
Split convert.cpp into smaller pieces
2018-02-06 20:02:51 +08:00
{
Merge pull request #12411 from vpisarev:wide_convert * rewrote Mat::convertTo() and convertScaleAbs() to wide universal intrinsics; added always-available and SIMD-optimized FP16<=>FP32 conversion * fixed compile warnings * fix some more compile errors * slightly relaxed accuracy threshold for int->float conversion (since we now do it using single-precision arithmetics, not double-precision) * fixed compile errors on iOS, Android and in the baseline C++ version (intrin_cpp.hpp) * trying to fix ARM-neon builds * trying to fix ARM-neon builds * trying to fix ARM-neon builds * trying to fix ARM-neon builds
2018-09-07 00:36:59 +08:00
if( j > size.width - VECSZ )
Split convert.cpp into smaller pieces
2018-02-06 20:02:51 +08:00
{
Merge pull request #12411 from vpisarev:wide_convert * rewrote Mat::convertTo() and convertScaleAbs() to wide universal intrinsics; added always-available and SIMD-optimized FP16<=>FP32 conversion * fixed compile warnings * fix some more compile errors * slightly relaxed accuracy threshold for int->float conversion (since we now do it using single-precision arithmetics, not double-precision) * fixed compile errors on iOS, Android and in the baseline C++ version (intrin_cpp.hpp) * trying to fix ARM-neon builds * trying to fix ARM-neon builds * trying to fix ARM-neon builds * trying to fix ARM-neon builds
2018-09-07 00:36:59 +08:00
if( j == 0 || src == (_Ts*)dst )
break;
j = size.width - VECSZ;
Split convert.cpp into smaller pieces
2018-02-06 20:02:51 +08:00
}
Merge pull request #12411 from vpisarev:wide_convert * rewrote Mat::convertTo() and convertScaleAbs() to wide universal intrinsics; added always-available and SIMD-optimized FP16<=>FP32 conversion * fixed compile warnings * fix some more compile errors * slightly relaxed accuracy threshold for int->float conversion (since we now do it using single-precision arithmetics, not double-precision) * fixed compile errors on iOS, Android and in the baseline C++ version (intrin_cpp.hpp) * trying to fix ARM-neon builds * trying to fix ARM-neon builds * trying to fix ARM-neon builds * trying to fix ARM-neon builds
2018-09-07 00:36:59 +08:00
v_float32 v0, v1;
vx_load_pair_as(src + j, v0, v1);
v0 = v_fma(v0, va, vb);
v1 = v_fma(v1, va, vb);
v_store_pair_as(dst + j, v_abs(v0), v_abs(v1));
Split convert.cpp into smaller pieces
2018-02-06 20:02:51 +08:00
}
#endif
Merge pull request #12411 from vpisarev:wide_convert * rewrote Mat::convertTo() and convertScaleAbs() to wide universal intrinsics; added always-available and SIMD-optimized FP16<=>FP32 conversion * fixed compile warnings * fix some more compile errors * slightly relaxed accuracy threshold for int->float conversion (since we now do it using single-precision arithmetics, not double-precision) * fixed compile errors on iOS, Android and in the baseline C++ version (intrin_cpp.hpp) * trying to fix ARM-neon builds * trying to fix ARM-neon builds * trying to fix ARM-neon builds * trying to fix ARM-neon builds
2018-09-07 00:36:59 +08:00
for( ; j < size.width; j++ )
dst[j] = saturate_cast<_Td>(std::abs(src[j]*a + b));
Split convert.cpp into smaller pieces
2018-02-06 20:02:51 +08:00
}
Merge pull request #12411 from vpisarev:wide_convert * rewrote Mat::convertTo() and convertScaleAbs() to wide universal intrinsics; added always-available and SIMD-optimized FP16<=>FP32 conversion * fixed compile warnings * fix some more compile errors * slightly relaxed accuracy threshold for int->float conversion (since we now do it using single-precision arithmetics, not double-precision) * fixed compile errors on iOS, Android and in the baseline C++ version (intrin_cpp.hpp) * trying to fix ARM-neon builds * trying to fix ARM-neon builds * trying to fix ARM-neon builds * trying to fix ARM-neon builds
2018-09-07 00:36:59 +08:00
}
Split convert.cpp into smaller pieces
2018-02-06 20:02:51 +08:00
Merge pull request #12411 from vpisarev:wide_convert * rewrote Mat::convertTo() and convertScaleAbs() to wide universal intrinsics; added always-available and SIMD-optimized FP16<=>FP32 conversion * fixed compile warnings * fix some more compile errors * slightly relaxed accuracy threshold for int->float conversion (since we now do it using single-precision arithmetics, not double-precision) * fixed compile errors on iOS, Android and in the baseline C++ version (intrin_cpp.hpp) * trying to fix ARM-neon builds * trying to fix ARM-neon builds * trying to fix ARM-neon builds * trying to fix ARM-neon builds
2018-09-07 00:36:59 +08:00
// variant for convrsions 16f <-> ... w/o unrolling
template<typename _Ts, typename _Td> inline void
cvtabs1_32f( const _Ts* src, size_t sstep, _Td* dst, size_t dstep,
Size size, float a, float b )
Split convert.cpp into smaller pieces
2018-02-06 20:02:51 +08:00
{
Merge pull request #12411 from vpisarev:wide_convert * rewrote Mat::convertTo() and convertScaleAbs() to wide universal intrinsics; added always-available and SIMD-optimized FP16<=>FP32 conversion * fixed compile warnings * fix some more compile errors * slightly relaxed accuracy threshold for int->float conversion (since we now do it using single-precision arithmetics, not double-precision) * fixed compile errors on iOS, Android and in the baseline C++ version (intrin_cpp.hpp) * trying to fix ARM-neon builds * trying to fix ARM-neon builds * trying to fix ARM-neon builds * trying to fix ARM-neon builds
2018-09-07 00:36:59 +08:00
#if CV_SIMD
v_float32 va = vx_setall_f32(a), vb = vx_setall_f32(b);
const int VECSZ = v_float32::nlanes*2;
Split convert.cpp into smaller pieces
2018-02-06 20:02:51 +08:00
#endif
Merge pull request #12411 from vpisarev:wide_convert * rewrote Mat::convertTo() and convertScaleAbs() to wide universal intrinsics; added always-available and SIMD-optimized FP16<=>FP32 conversion * fixed compile warnings * fix some more compile errors * slightly relaxed accuracy threshold for int->float conversion (since we now do it using single-precision arithmetics, not double-precision) * fixed compile errors on iOS, Android and in the baseline C++ version (intrin_cpp.hpp) * trying to fix ARM-neon builds * trying to fix ARM-neon builds * trying to fix ARM-neon builds * trying to fix ARM-neon builds
2018-09-07 00:36:59 +08:00
sstep /= sizeof(src[0]);
dstep /= sizeof(dst[0]);
Split convert.cpp into smaller pieces
2018-02-06 20:02:51 +08:00
Merge pull request #12411 from vpisarev:wide_convert * rewrote Mat::convertTo() and convertScaleAbs() to wide universal intrinsics; added always-available and SIMD-optimized FP16<=>FP32 conversion * fixed compile warnings * fix some more compile errors * slightly relaxed accuracy threshold for int->float conversion (since we now do it using single-precision arithmetics, not double-precision) * fixed compile errors on iOS, Android and in the baseline C++ version (intrin_cpp.hpp) * trying to fix ARM-neon builds * trying to fix ARM-neon builds * trying to fix ARM-neon builds * trying to fix ARM-neon builds
2018-09-07 00:36:59 +08:00
for( int i = 0; i < size.height; i++, src += sstep, dst += dstep )
Split convert.cpp into smaller pieces
2018-02-06 20:02:51 +08:00
{
Merge pull request #12411 from vpisarev:wide_convert * rewrote Mat::convertTo() and convertScaleAbs() to wide universal intrinsics; added always-available and SIMD-optimized FP16<=>FP32 conversion * fixed compile warnings * fix some more compile errors * slightly relaxed accuracy threshold for int->float conversion (since we now do it using single-precision arithmetics, not double-precision) * fixed compile errors on iOS, Android and in the baseline C++ version (intrin_cpp.hpp) * trying to fix ARM-neon builds * trying to fix ARM-neon builds * trying to fix ARM-neon builds * trying to fix ARM-neon builds
2018-09-07 00:36:59 +08:00
int j = 0;
#if CV_SIMD
for( ; j < size.width; j += VECSZ )
Split convert.cpp into smaller pieces
2018-02-06 20:02:51 +08:00
{
Merge pull request #12411 from vpisarev:wide_convert * rewrote Mat::convertTo() and convertScaleAbs() to wide universal intrinsics; added always-available and SIMD-optimized FP16<=>FP32 conversion * fixed compile warnings * fix some more compile errors * slightly relaxed accuracy threshold for int->float conversion (since we now do it using single-precision arithmetics, not double-precision) * fixed compile errors on iOS, Android and in the baseline C++ version (intrin_cpp.hpp) * trying to fix ARM-neon builds * trying to fix ARM-neon builds * trying to fix ARM-neon builds * trying to fix ARM-neon builds
2018-09-07 00:36:59 +08:00
if( j > size.width - VECSZ )
Split convert.cpp into smaller pieces
2018-02-06 20:02:51 +08:00
{
Merge pull request #12411 from vpisarev:wide_convert * rewrote Mat::convertTo() and convertScaleAbs() to wide universal intrinsics; added always-available and SIMD-optimized FP16<=>FP32 conversion * fixed compile warnings * fix some more compile errors * slightly relaxed accuracy threshold for int->float conversion (since we now do it using single-precision arithmetics, not double-precision) * fixed compile errors on iOS, Android and in the baseline C++ version (intrin_cpp.hpp) * trying to fix ARM-neon builds * trying to fix ARM-neon builds * trying to fix ARM-neon builds * trying to fix ARM-neon builds
2018-09-07 00:36:59 +08:00
if( j == 0 || src == (_Ts*)dst )
break;
j = size.width - VECSZ;
Split convert.cpp into smaller pieces
2018-02-06 20:02:51 +08:00
}
Merge pull request #12411 from vpisarev:wide_convert * rewrote Mat::convertTo() and convertScaleAbs() to wide universal intrinsics; added always-available and SIMD-optimized FP16<=>FP32 conversion * fixed compile warnings * fix some more compile errors * slightly relaxed accuracy threshold for int->float conversion (since we now do it using single-precision arithmetics, not double-precision) * fixed compile errors on iOS, Android and in the baseline C++ version (intrin_cpp.hpp) * trying to fix ARM-neon builds * trying to fix ARM-neon builds * trying to fix ARM-neon builds * trying to fix ARM-neon builds
2018-09-07 00:36:59 +08:00
v_float32 v0;
vx_load_as(src + j, v0);
v0 = v_fma(v0, va, vb);
v_store_as(dst + j, v_abs(v0));
Split convert.cpp into smaller pieces
2018-02-06 20:02:51 +08:00
}
Merge pull request #12411 from vpisarev:wide_convert * rewrote Mat::convertTo() and convertScaleAbs() to wide universal intrinsics; added always-available and SIMD-optimized FP16<=>FP32 conversion * fixed compile warnings * fix some more compile errors * slightly relaxed accuracy threshold for int->float conversion (since we now do it using single-precision arithmetics, not double-precision) * fixed compile errors on iOS, Android and in the baseline C++ version (intrin_cpp.hpp) * trying to fix ARM-neon builds * trying to fix ARM-neon builds * trying to fix ARM-neon builds * trying to fix ARM-neon builds
2018-09-07 00:36:59 +08:00
#endif
for( ; j < size.width; j++ )
dst[j] = saturate_cast<_Td>(src[j]*a + b);
Split convert.cpp into smaller pieces
2018-02-06 20:02:51 +08:00
}
}
Merge pull request #12411 from vpisarev:wide_convert * rewrote Mat::convertTo() and convertScaleAbs() to wide universal intrinsics; added always-available and SIMD-optimized FP16<=>FP32 conversion * fixed compile warnings * fix some more compile errors * slightly relaxed accuracy threshold for int->float conversion (since we now do it using single-precision arithmetics, not double-precision) * fixed compile errors on iOS, Android and in the baseline C++ version (intrin_cpp.hpp) * trying to fix ARM-neon builds * trying to fix ARM-neon builds * trying to fix ARM-neon builds * trying to fix ARM-neon builds
2018-09-07 00:36:59 +08:00
template<typename _Ts, typename _Td> inline void
cvt_32f( const _Ts* src, size_t sstep, _Td* dst, size_t dstep,
Size size, float a, float b )
Split convert.cpp into smaller pieces
2018-02-06 20:02:51 +08:00
{
Merge pull request #12411 from vpisarev:wide_convert * rewrote Mat::convertTo() and convertScaleAbs() to wide universal intrinsics; added always-available and SIMD-optimized FP16<=>FP32 conversion * fixed compile warnings * fix some more compile errors * slightly relaxed accuracy threshold for int->float conversion (since we now do it using single-precision arithmetics, not double-precision) * fixed compile errors on iOS, Android and in the baseline C++ version (intrin_cpp.hpp) * trying to fix ARM-neon builds * trying to fix ARM-neon builds * trying to fix ARM-neon builds * trying to fix ARM-neon builds
2018-09-07 00:36:59 +08:00
#if CV_SIMD
v_float32 va = vx_setall_f32(a), vb = vx_setall_f32(b);
const int VECSZ = v_float32::nlanes*2;
Split convert.cpp into smaller pieces
2018-02-06 20:02:51 +08:00
#endif
Merge pull request #12411 from vpisarev:wide_convert * rewrote Mat::convertTo() and convertScaleAbs() to wide universal intrinsics; added always-available and SIMD-optimized FP16<=>FP32 conversion * fixed compile warnings * fix some more compile errors * slightly relaxed accuracy threshold for int->float conversion (since we now do it using single-precision arithmetics, not double-precision) * fixed compile errors on iOS, Android and in the baseline C++ version (intrin_cpp.hpp) * trying to fix ARM-neon builds * trying to fix ARM-neon builds * trying to fix ARM-neon builds * trying to fix ARM-neon builds
2018-09-07 00:36:59 +08:00
sstep /= sizeof(src[0]);
dstep /= sizeof(dst[0]);
Split convert.cpp into smaller pieces
2018-02-06 20:02:51 +08:00
Merge pull request #12411 from vpisarev:wide_convert * rewrote Mat::convertTo() and convertScaleAbs() to wide universal intrinsics; added always-available and SIMD-optimized FP16<=>FP32 conversion * fixed compile warnings * fix some more compile errors * slightly relaxed accuracy threshold for int->float conversion (since we now do it using single-precision arithmetics, not double-precision) * fixed compile errors on iOS, Android and in the baseline C++ version (intrin_cpp.hpp) * trying to fix ARM-neon builds * trying to fix ARM-neon builds * trying to fix ARM-neon builds * trying to fix ARM-neon builds
2018-09-07 00:36:59 +08:00
for( int i = 0; i < size.height; i++, src += sstep, dst += dstep )
Split convert.cpp into smaller pieces
2018-02-06 20:02:51 +08:00
{
Merge pull request #12411 from vpisarev:wide_convert * rewrote Mat::convertTo() and convertScaleAbs() to wide universal intrinsics; added always-available and SIMD-optimized FP16<=>FP32 conversion * fixed compile warnings * fix some more compile errors * slightly relaxed accuracy threshold for int->float conversion (since we now do it using single-precision arithmetics, not double-precision) * fixed compile errors on iOS, Android and in the baseline C++ version (intrin_cpp.hpp) * trying to fix ARM-neon builds * trying to fix ARM-neon builds * trying to fix ARM-neon builds * trying to fix ARM-neon builds
2018-09-07 00:36:59 +08:00
int j = 0;
#if CV_SIMD
for( ; j < size.width; j += VECSZ )
Split convert.cpp into smaller pieces
2018-02-06 20:02:51 +08:00
{
Merge pull request #12411 from vpisarev:wide_convert * rewrote Mat::convertTo() and convertScaleAbs() to wide universal intrinsics; added always-available and SIMD-optimized FP16<=>FP32 conversion * fixed compile warnings * fix some more compile errors * slightly relaxed accuracy threshold for int->float conversion (since we now do it using single-precision arithmetics, not double-precision) * fixed compile errors on iOS, Android and in the baseline C++ version (intrin_cpp.hpp) * trying to fix ARM-neon builds * trying to fix ARM-neon builds * trying to fix ARM-neon builds * trying to fix ARM-neon builds
2018-09-07 00:36:59 +08:00
if( j > size.width - VECSZ )
Split convert.cpp into smaller pieces
2018-02-06 20:02:51 +08:00
{
Merge pull request #12411 from vpisarev:wide_convert * rewrote Mat::convertTo() and convertScaleAbs() to wide universal intrinsics; added always-available and SIMD-optimized FP16<=>FP32 conversion * fixed compile warnings * fix some more compile errors * slightly relaxed accuracy threshold for int->float conversion (since we now do it using single-precision arithmetics, not double-precision) * fixed compile errors on iOS, Android and in the baseline C++ version (intrin_cpp.hpp) * trying to fix ARM-neon builds * trying to fix ARM-neon builds * trying to fix ARM-neon builds * trying to fix ARM-neon builds
2018-09-07 00:36:59 +08:00
if( j == 0 || src == (_Ts*)dst )
break;
j = size.width - VECSZ;
Split convert.cpp into smaller pieces
2018-02-06 20:02:51 +08:00
}
Merge pull request #12411 from vpisarev:wide_convert * rewrote Mat::convertTo() and convertScaleAbs() to wide universal intrinsics; added always-available and SIMD-optimized FP16<=>FP32 conversion * fixed compile warnings * fix some more compile errors * slightly relaxed accuracy threshold for int->float conversion (since we now do it using single-precision arithmetics, not double-precision) * fixed compile errors on iOS, Android and in the baseline C++ version (intrin_cpp.hpp) * trying to fix ARM-neon builds * trying to fix ARM-neon builds * trying to fix ARM-neon builds * trying to fix ARM-neon builds
2018-09-07 00:36:59 +08:00
v_float32 v0, v1;
vx_load_pair_as(src + j, v0, v1);
v0 = v_fma(v0, va, vb);
v1 = v_fma(v1, va, vb);
v_store_pair_as(dst + j, v0, v1);
Split convert.cpp into smaller pieces
2018-02-06 20:02:51 +08:00
}
Merge pull request #12411 from vpisarev:wide_convert * rewrote Mat::convertTo() and convertScaleAbs() to wide universal intrinsics; added always-available and SIMD-optimized FP16<=>FP32 conversion * fixed compile warnings * fix some more compile errors * slightly relaxed accuracy threshold for int->float conversion (since we now do it using single-precision arithmetics, not double-precision) * fixed compile errors on iOS, Android and in the baseline C++ version (intrin_cpp.hpp) * trying to fix ARM-neon builds * trying to fix ARM-neon builds * trying to fix ARM-neon builds * trying to fix ARM-neon builds
2018-09-07 00:36:59 +08:00
#endif
for( ; j < size.width; j++ )
dst[j] = saturate_cast<_Td>(src[j]*a + b);
Split convert.cpp into smaller pieces
2018-02-06 20:02:51 +08:00
}
Merge pull request #12411 from vpisarev:wide_convert * rewrote Mat::convertTo() and convertScaleAbs() to wide universal intrinsics; added always-available and SIMD-optimized FP16<=>FP32 conversion * fixed compile warnings * fix some more compile errors * slightly relaxed accuracy threshold for int->float conversion (since we now do it using single-precision arithmetics, not double-precision) * fixed compile errors on iOS, Android and in the baseline C++ version (intrin_cpp.hpp) * trying to fix ARM-neon builds * trying to fix ARM-neon builds * trying to fix ARM-neon builds * trying to fix ARM-neon builds
2018-09-07 00:36:59 +08:00
}
Split convert.cpp into smaller pieces
2018-02-06 20:02:51 +08:00
Merge pull request #12411 from vpisarev:wide_convert * rewrote Mat::convertTo() and convertScaleAbs() to wide universal intrinsics; added always-available and SIMD-optimized FP16<=>FP32 conversion * fixed compile warnings * fix some more compile errors * slightly relaxed accuracy threshold for int->float conversion (since we now do it using single-precision arithmetics, not double-precision) * fixed compile errors on iOS, Android and in the baseline C++ version (intrin_cpp.hpp) * trying to fix ARM-neon builds * trying to fix ARM-neon builds * trying to fix ARM-neon builds * trying to fix ARM-neon builds
2018-09-07 00:36:59 +08:00
// variant for convrsions 16f <-> ... w/o unrolling
template<typename _Ts, typename _Td> inline void
cvt1_32f( const _Ts* src, size_t sstep, _Td* dst, size_t dstep,
Size size, float a, float b )
Split convert.cpp into smaller pieces
2018-02-06 20:02:51 +08:00
{
Merge pull request #12411 from vpisarev:wide_convert * rewrote Mat::convertTo() and convertScaleAbs() to wide universal intrinsics; added always-available and SIMD-optimized FP16<=>FP32 conversion * fixed compile warnings * fix some more compile errors * slightly relaxed accuracy threshold for int->float conversion (since we now do it using single-precision arithmetics, not double-precision) * fixed compile errors on iOS, Android and in the baseline C++ version (intrin_cpp.hpp) * trying to fix ARM-neon builds * trying to fix ARM-neon builds * trying to fix ARM-neon builds * trying to fix ARM-neon builds
2018-09-07 00:36:59 +08:00
#if CV_SIMD
v_float32 va = vx_setall_f32(a), vb = vx_setall_f32(b);
const int VECSZ = v_float32::nlanes;
Split convert.cpp into smaller pieces
2018-02-06 20:02:51 +08:00
#endif
sstep /= sizeof(src[0]);
dstep /= sizeof(dst[0]);
Merge pull request #12411 from vpisarev:wide_convert * rewrote Mat::convertTo() and convertScaleAbs() to wide universal intrinsics; added always-available and SIMD-optimized FP16<=>FP32 conversion * fixed compile warnings * fix some more compile errors * slightly relaxed accuracy threshold for int->float conversion (since we now do it using single-precision arithmetics, not double-precision) * fixed compile errors on iOS, Android and in the baseline C++ version (intrin_cpp.hpp) * trying to fix ARM-neon builds * trying to fix ARM-neon builds * trying to fix ARM-neon builds * trying to fix ARM-neon builds
2018-09-07 00:36:59 +08:00
for( int i = 0; i < size.height; i++, src += sstep, dst += dstep )
Split convert.cpp into smaller pieces
2018-02-06 20:02:51 +08:00
{
Merge pull request #12411 from vpisarev:wide_convert * rewrote Mat::convertTo() and convertScaleAbs() to wide universal intrinsics; added always-available and SIMD-optimized FP16<=>FP32 conversion * fixed compile warnings * fix some more compile errors * slightly relaxed accuracy threshold for int->float conversion (since we now do it using single-precision arithmetics, not double-precision) * fixed compile errors on iOS, Android and in the baseline C++ version (intrin_cpp.hpp) * trying to fix ARM-neon builds * trying to fix ARM-neon builds * trying to fix ARM-neon builds * trying to fix ARM-neon builds
2018-09-07 00:36:59 +08:00
int j = 0;
#if CV_SIMD
for( ; j < size.width; j += VECSZ )
Split convert.cpp into smaller pieces
2018-02-06 20:02:51 +08:00
{
Merge pull request #12411 from vpisarev:wide_convert * rewrote Mat::convertTo() and convertScaleAbs() to wide universal intrinsics; added always-available and SIMD-optimized FP16<=>FP32 conversion * fixed compile warnings * fix some more compile errors * slightly relaxed accuracy threshold for int->float conversion (since we now do it using single-precision arithmetics, not double-precision) * fixed compile errors on iOS, Android and in the baseline C++ version (intrin_cpp.hpp) * trying to fix ARM-neon builds * trying to fix ARM-neon builds * trying to fix ARM-neon builds * trying to fix ARM-neon builds
2018-09-07 00:36:59 +08:00
if( j > size.width - VECSZ )
{
if( j == 0 || src == (_Ts*)dst )
break;
j = size.width - VECSZ;
}
v_float32 v0;
vx_load_as(src + j, v0);
v0 = v_fma(v0, va, vb);
v_store_as(dst + j, v0);
Split convert.cpp into smaller pieces
2018-02-06 20:02:51 +08:00
}
Merge pull request #12411 from vpisarev:wide_convert * rewrote Mat::convertTo() and convertScaleAbs() to wide universal intrinsics; added always-available and SIMD-optimized FP16<=>FP32 conversion * fixed compile warnings * fix some more compile errors * slightly relaxed accuracy threshold for int->float conversion (since we now do it using single-precision arithmetics, not double-precision) * fixed compile errors on iOS, Android and in the baseline C++ version (intrin_cpp.hpp) * trying to fix ARM-neon builds * trying to fix ARM-neon builds * trying to fix ARM-neon builds * trying to fix ARM-neon builds
2018-09-07 00:36:59 +08:00
#endif
for( ; j < size.width; j++ )
dst[j] = saturate_cast<_Td>(src[j]*a + b);
Split convert.cpp into smaller pieces
2018-02-06 20:02:51 +08:00
}
}
Merge pull request #12411 from vpisarev:wide_convert * rewrote Mat::convertTo() and convertScaleAbs() to wide universal intrinsics; added always-available and SIMD-optimized FP16<=>FP32 conversion * fixed compile warnings * fix some more compile errors * slightly relaxed accuracy threshold for int->float conversion (since we now do it using single-precision arithmetics, not double-precision) * fixed compile errors on iOS, Android and in the baseline C++ version (intrin_cpp.hpp) * trying to fix ARM-neon builds * trying to fix ARM-neon builds * trying to fix ARM-neon builds * trying to fix ARM-neon builds
2018-09-07 00:36:59 +08:00
template<typename _Ts, typename _Td> inline void
cvt_64f( const _Ts* src, size_t sstep, _Td* dst, size_t dstep,
Size size, double a, double b )
Split convert.cpp into smaller pieces
2018-02-06 20:02:51 +08:00
{
Merge pull request #12411 from vpisarev:wide_convert * rewrote Mat::convertTo() and convertScaleAbs() to wide universal intrinsics; added always-available and SIMD-optimized FP16<=>FP32 conversion * fixed compile warnings * fix some more compile errors * slightly relaxed accuracy threshold for int->float conversion (since we now do it using single-precision arithmetics, not double-precision) * fixed compile errors on iOS, Android and in the baseline C++ version (intrin_cpp.hpp) * trying to fix ARM-neon builds * trying to fix ARM-neon builds * trying to fix ARM-neon builds * trying to fix ARM-neon builds
2018-09-07 00:36:59 +08:00
#if CV_SIMD_64F
v_float64 va = vx_setall_f64(a), vb = vx_setall_f64(b);
const int VECSZ = v_float64::nlanes*2;
#endif
Split convert.cpp into smaller pieces
2018-02-06 20:02:51 +08:00
sstep /= sizeof(src[0]);
dstep /= sizeof(dst[0]);
Merge pull request #12411 from vpisarev:wide_convert * rewrote Mat::convertTo() and convertScaleAbs() to wide universal intrinsics; added always-available and SIMD-optimized FP16<=>FP32 conversion * fixed compile warnings * fix some more compile errors * slightly relaxed accuracy threshold for int->float conversion (since we now do it using single-precision arithmetics, not double-precision) * fixed compile errors on iOS, Android and in the baseline C++ version (intrin_cpp.hpp) * trying to fix ARM-neon builds * trying to fix ARM-neon builds * trying to fix ARM-neon builds * trying to fix ARM-neon builds
2018-09-07 00:36:59 +08:00
for( int i = 0; i < size.height; i++, src += sstep, dst += dstep )
Split convert.cpp into smaller pieces
2018-02-06 20:02:51 +08:00
{
Merge pull request #12411 from vpisarev:wide_convert * rewrote Mat::convertTo() and convertScaleAbs() to wide universal intrinsics; added always-available and SIMD-optimized FP16<=>FP32 conversion * fixed compile warnings * fix some more compile errors * slightly relaxed accuracy threshold for int->float conversion (since we now do it using single-precision arithmetics, not double-precision) * fixed compile errors on iOS, Android and in the baseline C++ version (intrin_cpp.hpp) * trying to fix ARM-neon builds * trying to fix ARM-neon builds * trying to fix ARM-neon builds * trying to fix ARM-neon builds
2018-09-07 00:36:59 +08:00
int j = 0;
#if CV_SIMD_64F
for( ; j < size.width; j += VECSZ )
Split convert.cpp into smaller pieces
2018-02-06 20:02:51 +08:00
{
Merge pull request #12411 from vpisarev:wide_convert * rewrote Mat::convertTo() and convertScaleAbs() to wide universal intrinsics; added always-available and SIMD-optimized FP16<=>FP32 conversion * fixed compile warnings * fix some more compile errors * slightly relaxed accuracy threshold for int->float conversion (since we now do it using single-precision arithmetics, not double-precision) * fixed compile errors on iOS, Android and in the baseline C++ version (intrin_cpp.hpp) * trying to fix ARM-neon builds * trying to fix ARM-neon builds * trying to fix ARM-neon builds * trying to fix ARM-neon builds
2018-09-07 00:36:59 +08:00
if( j > size.width - VECSZ )
Split convert.cpp into smaller pieces
2018-02-06 20:02:51 +08:00
{
Merge pull request #12411 from vpisarev:wide_convert * rewrote Mat::convertTo() and convertScaleAbs() to wide universal intrinsics; added always-available and SIMD-optimized FP16<=>FP32 conversion * fixed compile warnings * fix some more compile errors * slightly relaxed accuracy threshold for int->float conversion (since we now do it using single-precision arithmetics, not double-precision) * fixed compile errors on iOS, Android and in the baseline C++ version (intrin_cpp.hpp) * trying to fix ARM-neon builds * trying to fix ARM-neon builds * trying to fix ARM-neon builds * trying to fix ARM-neon builds
2018-09-07 00:36:59 +08:00
if( j == 0 || src == (_Ts*)dst )
break;
j = size.width - VECSZ;
Split convert.cpp into smaller pieces
2018-02-06 20:02:51 +08:00
}
Merge pull request #12411 from vpisarev:wide_convert * rewrote Mat::convertTo() and convertScaleAbs() to wide universal intrinsics; added always-available and SIMD-optimized FP16<=>FP32 conversion * fixed compile warnings * fix some more compile errors * slightly relaxed accuracy threshold for int->float conversion (since we now do it using single-precision arithmetics, not double-precision) * fixed compile errors on iOS, Android and in the baseline C++ version (intrin_cpp.hpp) * trying to fix ARM-neon builds * trying to fix ARM-neon builds * trying to fix ARM-neon builds * trying to fix ARM-neon builds
2018-09-07 00:36:59 +08:00
v_float64 v0, v1;
vx_load_pair_as(src + j, v0, v1);
v0 = v_fma(v0, va, vb);
v1 = v_fma(v1, va, vb);
v_store_pair_as(dst + j, v0, v1);
Split convert.cpp into smaller pieces
2018-02-06 20:02:51 +08:00
}
Merge pull request #12411 from vpisarev:wide_convert * rewrote Mat::convertTo() and convertScaleAbs() to wide universal intrinsics; added always-available and SIMD-optimized FP16<=>FP32 conversion * fixed compile warnings * fix some more compile errors * slightly relaxed accuracy threshold for int->float conversion (since we now do it using single-precision arithmetics, not double-precision) * fixed compile errors on iOS, Android and in the baseline C++ version (intrin_cpp.hpp) * trying to fix ARM-neon builds * trying to fix ARM-neon builds * trying to fix ARM-neon builds * trying to fix ARM-neon builds
2018-09-07 00:36:59 +08:00
#endif
for( ; j < size.width; j++ )
dst[j] = saturate_cast<_Td>(src[j]*a + b);
Split convert.cpp into smaller pieces
2018-02-06 20:02:51 +08:00
}
}
//==================================================================================================
Merge pull request #12411 from vpisarev:wide_convert * rewrote Mat::convertTo() and convertScaleAbs() to wide universal intrinsics; added always-available and SIMD-optimized FP16<=>FP32 conversion * fixed compile warnings * fix some more compile errors * slightly relaxed accuracy threshold for int->float conversion (since we now do it using single-precision arithmetics, not double-precision) * fixed compile errors on iOS, Android and in the baseline C++ version (intrin_cpp.hpp) * trying to fix ARM-neon builds * trying to fix ARM-neon builds * trying to fix ARM-neon builds * trying to fix ARM-neon builds
2018-09-07 00:36:59 +08:00
#define DEF_CVT_SCALE_ABS_FUNC(suffix, cvt, stype, dtype, wtype) \
core: avoid function type cast, make happy UBSAN oss-fuzz: https://bugs.chromium.org/p/oss-fuzz/issues/detail?id=14115
2019-06-11 15:06:29 +08:00
static void cvtScaleAbs##suffix( const uchar* src_, size_t sstep, const uchar*, size_t, \
uchar* dst_, size_t dstep, Size size, void* scale_) \
Split convert.cpp into smaller pieces
2018-02-06 20:02:51 +08:00
{ \
core: avoid function type cast, make happy UBSAN oss-fuzz: https://bugs.chromium.org/p/oss-fuzz/issues/detail?id=14115
2019-06-11 15:06:29 +08:00
const stype* src = (const stype*)src_; \
dtype* dst = (dtype*)dst_; \
double* scale = (double*)scale_; \
Merge pull request #12411 from vpisarev:wide_convert * rewrote Mat::convertTo() and convertScaleAbs() to wide universal intrinsics; added always-available and SIMD-optimized FP16<=>FP32 conversion * fixed compile warnings * fix some more compile errors * slightly relaxed accuracy threshold for int->float conversion (since we now do it using single-precision arithmetics, not double-precision) * fixed compile errors on iOS, Android and in the baseline C++ version (intrin_cpp.hpp) * trying to fix ARM-neon builds * trying to fix ARM-neon builds * trying to fix ARM-neon builds * trying to fix ARM-neon builds
2018-09-07 00:36:59 +08:00
cvt(src, sstep, dst, dstep, size, (wtype)scale[0], (wtype)scale[1]); \
Split convert.cpp into smaller pieces
2018-02-06 20:02:51 +08:00
}
Merge pull request #12411 from vpisarev:wide_convert * rewrote Mat::convertTo() and convertScaleAbs() to wide universal intrinsics; added always-available and SIMD-optimized FP16<=>FP32 conversion * fixed compile warnings * fix some more compile errors * slightly relaxed accuracy threshold for int->float conversion (since we now do it using single-precision arithmetics, not double-precision) * fixed compile errors on iOS, Android and in the baseline C++ version (intrin_cpp.hpp) * trying to fix ARM-neon builds * trying to fix ARM-neon builds * trying to fix ARM-neon builds * trying to fix ARM-neon builds
2018-09-07 00:36:59 +08:00
#define DEF_CVT_SCALE_FUNC(suffix, cvt, stype, dtype, wtype) \
core: avoid function type cast, make happy UBSAN oss-fuzz: https://bugs.chromium.org/p/oss-fuzz/issues/detail?id=14115
2019-06-11 15:06:29 +08:00
static void cvtScale##suffix( const uchar* src_, size_t sstep, const uchar*, size_t, \
uchar* dst_, size_t dstep, Size size, void* scale_) \
Split convert.cpp into smaller pieces
2018-02-06 20:02:51 +08:00
{ \
core: avoid function type cast, make happy UBSAN oss-fuzz: https://bugs.chromium.org/p/oss-fuzz/issues/detail?id=14115
2019-06-11 15:06:29 +08:00
const stype* src = (const stype*)src_; \
dtype* dst = (dtype*)dst_; \
double* scale = (double*)scale_; \
Merge pull request #12411 from vpisarev:wide_convert * rewrote Mat::convertTo() and convertScaleAbs() to wide universal intrinsics; added always-available and SIMD-optimized FP16<=>FP32 conversion * fixed compile warnings * fix some more compile errors * slightly relaxed accuracy threshold for int->float conversion (since we now do it using single-precision arithmetics, not double-precision) * fixed compile errors on iOS, Android and in the baseline C++ version (intrin_cpp.hpp) * trying to fix ARM-neon builds * trying to fix ARM-neon builds * trying to fix ARM-neon builds * trying to fix ARM-neon builds
2018-09-07 00:36:59 +08:00
cvt(src, sstep, dst, dstep, size, (wtype)scale[0], (wtype)scale[1]); \
Split convert.cpp into smaller pieces
2018-02-06 20:02:51 +08:00
}
Merge pull request #12411 from vpisarev:wide_convert * rewrote Mat::convertTo() and convertScaleAbs() to wide universal intrinsics; added always-available and SIMD-optimized FP16<=>FP32 conversion * fixed compile warnings * fix some more compile errors * slightly relaxed accuracy threshold for int->float conversion (since we now do it using single-precision arithmetics, not double-precision) * fixed compile errors on iOS, Android and in the baseline C++ version (intrin_cpp.hpp) * trying to fix ARM-neon builds * trying to fix ARM-neon builds * trying to fix ARM-neon builds * trying to fix ARM-neon builds
2018-09-07 00:36:59 +08:00
DEF_CVT_SCALE_ABS_FUNC(8u, cvtabs_32f, uchar, uchar, float)
DEF_CVT_SCALE_ABS_FUNC(8s8u, cvtabs_32f, schar, uchar, float)
DEF_CVT_SCALE_ABS_FUNC(16u8u, cvtabs_32f, ushort, uchar, float)
DEF_CVT_SCALE_ABS_FUNC(16s8u, cvtabs_32f, short, uchar, float)
DEF_CVT_SCALE_ABS_FUNC(32s8u, cvtabs_32f, int, uchar, float)
DEF_CVT_SCALE_ABS_FUNC(32f8u, cvtabs_32f, float, uchar, float)
DEF_CVT_SCALE_ABS_FUNC(64f8u, cvtabs_32f, double, uchar, float)
DEF_CVT_SCALE_FUNC(8u, cvt_32f, uchar, uchar, float)
DEF_CVT_SCALE_FUNC(8s8u, cvt_32f, schar, uchar, float)
DEF_CVT_SCALE_FUNC(16u8u, cvt_32f, ushort, uchar, float)
DEF_CVT_SCALE_FUNC(16s8u, cvt_32f, short, uchar, float)
DEF_CVT_SCALE_FUNC(32s8u, cvt_32f, int, uchar, float)
DEF_CVT_SCALE_FUNC(32f8u, cvt_32f, float, uchar, float)
DEF_CVT_SCALE_FUNC(64f8u, cvt_32f, double, uchar, float)
added basic support for CV_16F (the new datatype etc.) (#12463) * added basic support for CV_16F (the new datatype etc.). CV_USRTYPE1 is now equal to CV_16F, which may break some [rarely used] functionality. We'll see * fixed just introduced bug in norm; reverted errorneous changes in Torch importer (need to find a better solution) * addressed some issues found during the PR review * restored the patch to fix some perf test failures
2018-09-10 21:56:29 +08:00
DEF_CVT_SCALE_FUNC(16f8u, cvt_32f, float16_t, uchar, float)
Merge pull request #12411 from vpisarev:wide_convert * rewrote Mat::convertTo() and convertScaleAbs() to wide universal intrinsics; added always-available and SIMD-optimized FP16<=>FP32 conversion * fixed compile warnings * fix some more compile errors * slightly relaxed accuracy threshold for int->float conversion (since we now do it using single-precision arithmetics, not double-precision) * fixed compile errors on iOS, Android and in the baseline C++ version (intrin_cpp.hpp) * trying to fix ARM-neon builds * trying to fix ARM-neon builds * trying to fix ARM-neon builds * trying to fix ARM-neon builds
2018-09-07 00:36:59 +08:00
DEF_CVT_SCALE_FUNC(8u8s, cvt_32f, uchar, schar, float)
DEF_CVT_SCALE_FUNC(8s, cvt_32f, schar, schar, float)
DEF_CVT_SCALE_FUNC(16u8s, cvt_32f, ushort, schar, float)
DEF_CVT_SCALE_FUNC(16s8s, cvt_32f, short, schar, float)
DEF_CVT_SCALE_FUNC(32s8s, cvt_32f, int, schar, float)
DEF_CVT_SCALE_FUNC(32f8s, cvt_32f, float, schar, float)
DEF_CVT_SCALE_FUNC(64f8s, cvt_32f, double, schar, float)
added basic support for CV_16F (the new datatype etc.) (#12463) * added basic support for CV_16F (the new datatype etc.). CV_USRTYPE1 is now equal to CV_16F, which may break some [rarely used] functionality. We'll see * fixed just introduced bug in norm; reverted errorneous changes in Torch importer (need to find a better solution) * addressed some issues found during the PR review * restored the patch to fix some perf test failures
2018-09-10 21:56:29 +08:00
DEF_CVT_SCALE_FUNC(16f8s, cvt_32f, float16_t, schar, float)
Merge pull request #12411 from vpisarev:wide_convert * rewrote Mat::convertTo() and convertScaleAbs() to wide universal intrinsics; added always-available and SIMD-optimized FP16<=>FP32 conversion * fixed compile warnings * fix some more compile errors * slightly relaxed accuracy threshold for int->float conversion (since we now do it using single-precision arithmetics, not double-precision) * fixed compile errors on iOS, Android and in the baseline C++ version (intrin_cpp.hpp) * trying to fix ARM-neon builds * trying to fix ARM-neon builds * trying to fix ARM-neon builds * trying to fix ARM-neon builds
2018-09-07 00:36:59 +08:00
DEF_CVT_SCALE_FUNC(8u16u, cvt_32f, uchar, ushort, float)
DEF_CVT_SCALE_FUNC(8s16u, cvt_32f, schar, ushort, float)
DEF_CVT_SCALE_FUNC(16u, cvt_32f, ushort, ushort, float)
DEF_CVT_SCALE_FUNC(16s16u, cvt_32f, short, ushort, float)
DEF_CVT_SCALE_FUNC(32s16u, cvt_32f, int, ushort, float)
DEF_CVT_SCALE_FUNC(32f16u, cvt_32f, float, ushort, float)
DEF_CVT_SCALE_FUNC(64f16u, cvt_32f, double, ushort, float)
added basic support for CV_16F (the new datatype etc.) (#12463) * added basic support for CV_16F (the new datatype etc.). CV_USRTYPE1 is now equal to CV_16F, which may break some [rarely used] functionality. We'll see * fixed just introduced bug in norm; reverted errorneous changes in Torch importer (need to find a better solution) * addressed some issues found during the PR review * restored the patch to fix some perf test failures
2018-09-10 21:56:29 +08:00
DEF_CVT_SCALE_FUNC(16f16u, cvt1_32f, float16_t, ushort, float)
Merge pull request #12411 from vpisarev:wide_convert * rewrote Mat::convertTo() and convertScaleAbs() to wide universal intrinsics; added always-available and SIMD-optimized FP16<=>FP32 conversion * fixed compile warnings * fix some more compile errors * slightly relaxed accuracy threshold for int->float conversion (since we now do it using single-precision arithmetics, not double-precision) * fixed compile errors on iOS, Android and in the baseline C++ version (intrin_cpp.hpp) * trying to fix ARM-neon builds * trying to fix ARM-neon builds * trying to fix ARM-neon builds * trying to fix ARM-neon builds
2018-09-07 00:36:59 +08:00
DEF_CVT_SCALE_FUNC(8u16s, cvt_32f, uchar, short, float)
DEF_CVT_SCALE_FUNC(8s16s, cvt_32f, schar, short, float)
DEF_CVT_SCALE_FUNC(16u16s, cvt_32f, ushort, short, float)
DEF_CVT_SCALE_FUNC(16s, cvt_32f, short, short, float)
DEF_CVT_SCALE_FUNC(32s16s, cvt_32f, int, short, float)
DEF_CVT_SCALE_FUNC(32f16s, cvt_32f, float, short, float)
DEF_CVT_SCALE_FUNC(64f16s, cvt_32f, double, short, float)
added basic support for CV_16F (the new datatype etc.) (#12463) * added basic support for CV_16F (the new datatype etc.). CV_USRTYPE1 is now equal to CV_16F, which may break some [rarely used] functionality. We'll see * fixed just introduced bug in norm; reverted errorneous changes in Torch importer (need to find a better solution) * addressed some issues found during the PR review * restored the patch to fix some perf test failures
2018-09-10 21:56:29 +08:00
DEF_CVT_SCALE_FUNC(16f16s, cvt1_32f, float16_t, short, float)
Merge pull request #12411 from vpisarev:wide_convert * rewrote Mat::convertTo() and convertScaleAbs() to wide universal intrinsics; added always-available and SIMD-optimized FP16<=>FP32 conversion * fixed compile warnings * fix some more compile errors * slightly relaxed accuracy threshold for int->float conversion (since we now do it using single-precision arithmetics, not double-precision) * fixed compile errors on iOS, Android and in the baseline C++ version (intrin_cpp.hpp) * trying to fix ARM-neon builds * trying to fix ARM-neon builds * trying to fix ARM-neon builds * trying to fix ARM-neon builds
2018-09-07 00:36:59 +08:00
DEF_CVT_SCALE_FUNC(8u32s, cvt_32f, uchar, int, float)
DEF_CVT_SCALE_FUNC(8s32s, cvt_32f, schar, int, float)
DEF_CVT_SCALE_FUNC(16u32s, cvt_32f, ushort, int, float)
DEF_CVT_SCALE_FUNC(16s32s, cvt_32f, short, int, float)
DEF_CVT_SCALE_FUNC(32s, cvt_64f, int, int, double)
DEF_CVT_SCALE_FUNC(32f32s, cvt_32f, float, int, float)
DEF_CVT_SCALE_FUNC(64f32s, cvt_64f, double, int, double)
added basic support for CV_16F (the new datatype etc.) (#12463) * added basic support for CV_16F (the new datatype etc.). CV_USRTYPE1 is now equal to CV_16F, which may break some [rarely used] functionality. We'll see * fixed just introduced bug in norm; reverted errorneous changes in Torch importer (need to find a better solution) * addressed some issues found during the PR review * restored the patch to fix some perf test failures
2018-09-10 21:56:29 +08:00
DEF_CVT_SCALE_FUNC(16f32s, cvt1_32f, float16_t, int, float)
Merge pull request #12411 from vpisarev:wide_convert * rewrote Mat::convertTo() and convertScaleAbs() to wide universal intrinsics; added always-available and SIMD-optimized FP16<=>FP32 conversion * fixed compile warnings * fix some more compile errors * slightly relaxed accuracy threshold for int->float conversion (since we now do it using single-precision arithmetics, not double-precision) * fixed compile errors on iOS, Android and in the baseline C++ version (intrin_cpp.hpp) * trying to fix ARM-neon builds * trying to fix ARM-neon builds * trying to fix ARM-neon builds * trying to fix ARM-neon builds
2018-09-07 00:36:59 +08:00
DEF_CVT_SCALE_FUNC(8u32f, cvt_32f, uchar, float, float)
DEF_CVT_SCALE_FUNC(8s32f, cvt_32f, schar, float, float)
DEF_CVT_SCALE_FUNC(16u32f, cvt_32f, ushort, float, float)
DEF_CVT_SCALE_FUNC(16s32f, cvt_32f, short, float, float)
DEF_CVT_SCALE_FUNC(32s32f, cvt_32f, int, float, float)
DEF_CVT_SCALE_FUNC(32f, cvt_32f, float, float, float)
DEF_CVT_SCALE_FUNC(64f32f, cvt_64f, double, float, double)
added basic support for CV_16F (the new datatype etc.) (#12463) * added basic support for CV_16F (the new datatype etc.). CV_USRTYPE1 is now equal to CV_16F, which may break some [rarely used] functionality. We'll see * fixed just introduced bug in norm; reverted errorneous changes in Torch importer (need to find a better solution) * addressed some issues found during the PR review * restored the patch to fix some perf test failures
2018-09-10 21:56:29 +08:00
DEF_CVT_SCALE_FUNC(16f32f, cvt1_32f, float16_t, float, float)
Merge pull request #12411 from vpisarev:wide_convert * rewrote Mat::convertTo() and convertScaleAbs() to wide universal intrinsics; added always-available and SIMD-optimized FP16<=>FP32 conversion * fixed compile warnings * fix some more compile errors * slightly relaxed accuracy threshold for int->float conversion (since we now do it using single-precision arithmetics, not double-precision) * fixed compile errors on iOS, Android and in the baseline C++ version (intrin_cpp.hpp) * trying to fix ARM-neon builds * trying to fix ARM-neon builds * trying to fix ARM-neon builds * trying to fix ARM-neon builds
2018-09-07 00:36:59 +08:00
DEF_CVT_SCALE_FUNC(8u64f, cvt_64f, uchar, double, double)
DEF_CVT_SCALE_FUNC(8s64f, cvt_64f, schar, double, double)
DEF_CVT_SCALE_FUNC(16u64f, cvt_64f, ushort, double, double)
DEF_CVT_SCALE_FUNC(16s64f, cvt_64f, short, double, double)
DEF_CVT_SCALE_FUNC(32s64f, cvt_64f, int, double, double)
DEF_CVT_SCALE_FUNC(32f64f, cvt_64f, float, double, double)
DEF_CVT_SCALE_FUNC(64f, cvt_64f, double, double, double)
added basic support for CV_16F (the new datatype etc.) (#12463) * added basic support for CV_16F (the new datatype etc.). CV_USRTYPE1 is now equal to CV_16F, which may break some [rarely used] functionality. We'll see * fixed just introduced bug in norm; reverted errorneous changes in Torch importer (need to find a better solution) * addressed some issues found during the PR review * restored the patch to fix some perf test failures
2018-09-10 21:56:29 +08:00
DEF_CVT_SCALE_FUNC(16f64f, cvt_64f, float16_t, double, double)
Merge pull request #12411 from vpisarev:wide_convert * rewrote Mat::convertTo() and convertScaleAbs() to wide universal intrinsics; added always-available and SIMD-optimized FP16<=>FP32 conversion * fixed compile warnings * fix some more compile errors * slightly relaxed accuracy threshold for int->float conversion (since we now do it using single-precision arithmetics, not double-precision) * fixed compile errors on iOS, Android and in the baseline C++ version (intrin_cpp.hpp) * trying to fix ARM-neon builds * trying to fix ARM-neon builds * trying to fix ARM-neon builds * trying to fix ARM-neon builds
2018-09-07 00:36:59 +08:00
added basic support for CV_16F (the new datatype etc.) (#12463) * added basic support for CV_16F (the new datatype etc.). CV_USRTYPE1 is now equal to CV_16F, which may break some [rarely used] functionality. We'll see * fixed just introduced bug in norm; reverted errorneous changes in Torch importer (need to find a better solution) * addressed some issues found during the PR review * restored the patch to fix some perf test failures
2018-09-10 21:56:29 +08:00
DEF_CVT_SCALE_FUNC(8u16f, cvt1_32f, uchar, float16_t, float)
Merge pull request #12411 from vpisarev:wide_convert * rewrote Mat::convertTo() and convertScaleAbs() to wide universal intrinsics; added always-available and SIMD-optimized FP16<=>FP32 conversion * fixed compile warnings * fix some more compile errors * slightly relaxed accuracy threshold for int->float conversion (since we now do it using single-precision arithmetics, not double-precision) * fixed compile errors on iOS, Android and in the baseline C++ version (intrin_cpp.hpp) * trying to fix ARM-neon builds * trying to fix ARM-neon builds * trying to fix ARM-neon builds * trying to fix ARM-neon builds
2018-09-07 00:36:59 +08:00
DEF_CVT_SCALE_FUNC(8s16f, cvt1_32f, schar, float16_t, float)
DEF_CVT_SCALE_FUNC(16u16f, cvt1_32f, ushort, float16_t, float)
DEF_CVT_SCALE_FUNC(16s16f, cvt1_32f, short, float16_t, float)
DEF_CVT_SCALE_FUNC(32s16f, cvt1_32f, int, float16_t, float)
DEF_CVT_SCALE_FUNC(32f16f, cvt1_32f, float, float16_t, float)
DEF_CVT_SCALE_FUNC(64f16f, cvt_64f, double, float16_t, double)
added basic support for CV_16F (the new datatype etc.) (#12463) * added basic support for CV_16F (the new datatype etc.). CV_USRTYPE1 is now equal to CV_16F, which may break some [rarely used] functionality. We'll see * fixed just introduced bug in norm; reverted errorneous changes in Torch importer (need to find a better solution) * addressed some issues found during the PR review * restored the patch to fix some perf test failures
2018-09-10 21:56:29 +08:00
DEF_CVT_SCALE_FUNC(16f, cvt1_32f, float16_t, float16_t, float)
Split convert.cpp into smaller pieces
2018-02-06 20:02:51 +08:00
core: dispatch convert_scale
2019-02-08 22:56:25 +08:00
BinaryFunc getCvtScaleAbsFunc(int depth)
Split convert.cpp into smaller pieces
2018-02-06 20:02:51 +08:00
{
static BinaryFunc cvtScaleAbsTab[] =
{
(BinaryFunc)cvtScaleAbs8u, (BinaryFunc)cvtScaleAbs8s8u, (BinaryFunc)cvtScaleAbs16u8u,
(BinaryFunc)cvtScaleAbs16s8u, (BinaryFunc)cvtScaleAbs32s8u, (BinaryFunc)cvtScaleAbs32f8u,
(BinaryFunc)cvtScaleAbs64f8u, 0
};
return cvtScaleAbsTab[depth];
}
BinaryFunc getConvertScaleFunc(int sdepth, int ddepth)
{
static BinaryFunc cvtScaleTab[][8] =
{
{
(BinaryFunc)GET_OPTIMIZED(cvtScale8u), (BinaryFunc)GET_OPTIMIZED(cvtScale8s8u), (BinaryFunc)GET_OPTIMIZED(cvtScale16u8u),
(BinaryFunc)GET_OPTIMIZED(cvtScale16s8u), (BinaryFunc)GET_OPTIMIZED(cvtScale32s8u), (BinaryFunc)GET_OPTIMIZED(cvtScale32f8u),
added basic support for CV_16F (the new datatype etc.) (#12463) * added basic support for CV_16F (the new datatype etc.). CV_USRTYPE1 is now equal to CV_16F, which may break some [rarely used] functionality. We'll see * fixed just introduced bug in norm; reverted errorneous changes in Torch importer (need to find a better solution) * addressed some issues found during the PR review * restored the patch to fix some perf test failures
2018-09-10 21:56:29 +08:00
(BinaryFunc)cvtScale64f8u, (BinaryFunc)cvtScale16f8u
Split convert.cpp into smaller pieces
2018-02-06 20:02:51 +08:00
},
{
(BinaryFunc)GET_OPTIMIZED(cvtScale8u8s), (BinaryFunc)GET_OPTIMIZED(cvtScale8s), (BinaryFunc)GET_OPTIMIZED(cvtScale16u8s),
(BinaryFunc)GET_OPTIMIZED(cvtScale16s8s), (BinaryFunc)GET_OPTIMIZED(cvtScale32s8s), (BinaryFunc)GET_OPTIMIZED(cvtScale32f8s),
added basic support for CV_16F (the new datatype etc.) (#12463) * added basic support for CV_16F (the new datatype etc.). CV_USRTYPE1 is now equal to CV_16F, which may break some [rarely used] functionality. We'll see * fixed just introduced bug in norm; reverted errorneous changes in Torch importer (need to find a better solution) * addressed some issues found during the PR review * restored the patch to fix some perf test failures
2018-09-10 21:56:29 +08:00
(BinaryFunc)cvtScale64f8s, (BinaryFunc)cvtScale16f8s
Split convert.cpp into smaller pieces
2018-02-06 20:02:51 +08:00
},
{
(BinaryFunc)GET_OPTIMIZED(cvtScale8u16u), (BinaryFunc)GET_OPTIMIZED(cvtScale8s16u), (BinaryFunc)GET_OPTIMIZED(cvtScale16u),
(BinaryFunc)GET_OPTIMIZED(cvtScale16s16u), (BinaryFunc)GET_OPTIMIZED(cvtScale32s16u), (BinaryFunc)GET_OPTIMIZED(cvtScale32f16u),
added basic support for CV_16F (the new datatype etc.) (#12463) * added basic support for CV_16F (the new datatype etc.). CV_USRTYPE1 is now equal to CV_16F, which may break some [rarely used] functionality. We'll see * fixed just introduced bug in norm; reverted errorneous changes in Torch importer (need to find a better solution) * addressed some issues found during the PR review * restored the patch to fix some perf test failures
2018-09-10 21:56:29 +08:00
(BinaryFunc)cvtScale64f16u, (BinaryFunc)cvtScale16f16u
Split convert.cpp into smaller pieces
2018-02-06 20:02:51 +08:00
},
{
(BinaryFunc)GET_OPTIMIZED(cvtScale8u16s), (BinaryFunc)GET_OPTIMIZED(cvtScale8s16s), (BinaryFunc)GET_OPTIMIZED(cvtScale16u16s),
(BinaryFunc)GET_OPTIMIZED(cvtScale16s), (BinaryFunc)GET_OPTIMIZED(cvtScale32s16s), (BinaryFunc)GET_OPTIMIZED(cvtScale32f16s),
added basic support for CV_16F (the new datatype etc.) (#12463) * added basic support for CV_16F (the new datatype etc.). CV_USRTYPE1 is now equal to CV_16F, which may break some [rarely used] functionality. We'll see * fixed just introduced bug in norm; reverted errorneous changes in Torch importer (need to find a better solution) * addressed some issues found during the PR review * restored the patch to fix some perf test failures
2018-09-10 21:56:29 +08:00
(BinaryFunc)cvtScale64f16s, (BinaryFunc)cvtScale16f16s
Split convert.cpp into smaller pieces
2018-02-06 20:02:51 +08:00
},
{
(BinaryFunc)GET_OPTIMIZED(cvtScale8u32s), (BinaryFunc)GET_OPTIMIZED(cvtScale8s32s), (BinaryFunc)GET_OPTIMIZED(cvtScale16u32s),
(BinaryFunc)GET_OPTIMIZED(cvtScale16s32s), (BinaryFunc)GET_OPTIMIZED(cvtScale32s), (BinaryFunc)GET_OPTIMIZED(cvtScale32f32s),
added basic support for CV_16F (the new datatype etc.) (#12463) * added basic support for CV_16F (the new datatype etc.). CV_USRTYPE1 is now equal to CV_16F, which may break some [rarely used] functionality. We'll see * fixed just introduced bug in norm; reverted errorneous changes in Torch importer (need to find a better solution) * addressed some issues found during the PR review * restored the patch to fix some perf test failures
2018-09-10 21:56:29 +08:00
(BinaryFunc)cvtScale64f32s, (BinaryFunc)cvtScale16f32s
Split convert.cpp into smaller pieces
2018-02-06 20:02:51 +08:00
},
{
(BinaryFunc)GET_OPTIMIZED(cvtScale8u32f), (BinaryFunc)GET_OPTIMIZED(cvtScale8s32f), (BinaryFunc)GET_OPTIMIZED(cvtScale16u32f),
(BinaryFunc)GET_OPTIMIZED(cvtScale16s32f), (BinaryFunc)GET_OPTIMIZED(cvtScale32s32f), (BinaryFunc)GET_OPTIMIZED(cvtScale32f),
added basic support for CV_16F (the new datatype etc.) (#12463) * added basic support for CV_16F (the new datatype etc.). CV_USRTYPE1 is now equal to CV_16F, which may break some [rarely used] functionality. We'll see * fixed just introduced bug in norm; reverted errorneous changes in Torch importer (need to find a better solution) * addressed some issues found during the PR review * restored the patch to fix some perf test failures
2018-09-10 21:56:29 +08:00
(BinaryFunc)cvtScale64f32f, (BinaryFunc)cvtScale16f32f
Split convert.cpp into smaller pieces
2018-02-06 20:02:51 +08:00
},
{
(BinaryFunc)cvtScale8u64f, (BinaryFunc)cvtScale8s64f, (BinaryFunc)cvtScale16u64f,
(BinaryFunc)cvtScale16s64f, (BinaryFunc)cvtScale32s64f, (BinaryFunc)cvtScale32f64f,
added basic support for CV_16F (the new datatype etc.) (#12463) * added basic support for CV_16F (the new datatype etc.). CV_USRTYPE1 is now equal to CV_16F, which may break some [rarely used] functionality. We'll see * fixed just introduced bug in norm; reverted errorneous changes in Torch importer (need to find a better solution) * addressed some issues found during the PR review * restored the patch to fix some perf test failures
2018-09-10 21:56:29 +08:00
(BinaryFunc)cvtScale64f, (BinaryFunc)cvtScale16f64f
Split convert.cpp into smaller pieces
2018-02-06 20:02:51 +08:00
},
{
added basic support for CV_16F (the new datatype etc.) (#12463) * added basic support for CV_16F (the new datatype etc.). CV_USRTYPE1 is now equal to CV_16F, which may break some [rarely used] functionality. We'll see * fixed just introduced bug in norm; reverted errorneous changes in Torch importer (need to find a better solution) * addressed some issues found during the PR review * restored the patch to fix some perf test failures
2018-09-10 21:56:29 +08:00
(BinaryFunc)cvtScale8u16f, (BinaryFunc)cvtScale8s16f, (BinaryFunc)cvtScale16u16f,
Merge pull request #12411 from vpisarev:wide_convert * rewrote Mat::convertTo() and convertScaleAbs() to wide universal intrinsics; added always-available and SIMD-optimized FP16<=>FP32 conversion * fixed compile warnings * fix some more compile errors * slightly relaxed accuracy threshold for int->float conversion (since we now do it using single-precision arithmetics, not double-precision) * fixed compile errors on iOS, Android and in the baseline C++ version (intrin_cpp.hpp) * trying to fix ARM-neon builds * trying to fix ARM-neon builds * trying to fix ARM-neon builds * trying to fix ARM-neon builds
2018-09-07 00:36:59 +08:00
(BinaryFunc)cvtScale16s16f, (BinaryFunc)cvtScale32s16f, (BinaryFunc)cvtScale32f16f,
added basic support for CV_16F (the new datatype etc.) (#12463) * added basic support for CV_16F (the new datatype etc.). CV_USRTYPE1 is now equal to CV_16F, which may break some [rarely used] functionality. We'll see * fixed just introduced bug in norm; reverted errorneous changes in Torch importer (need to find a better solution) * addressed some issues found during the PR review * restored the patch to fix some perf test failures
2018-09-10 21:56:29 +08:00
(BinaryFunc)cvtScale64f16f, (BinaryFunc)cvtScale16f
Merge pull request #12411 from vpisarev:wide_convert * rewrote Mat::convertTo() and convertScaleAbs() to wide universal intrinsics; added always-available and SIMD-optimized FP16<=>FP32 conversion * fixed compile warnings * fix some more compile errors * slightly relaxed accuracy threshold for int->float conversion (since we now do it using single-precision arithmetics, not double-precision) * fixed compile errors on iOS, Android and in the baseline C++ version (intrin_cpp.hpp) * trying to fix ARM-neon builds * trying to fix ARM-neon builds * trying to fix ARM-neon builds * trying to fix ARM-neon builds
2018-09-07 00:36:59 +08:00
},
Split convert.cpp into smaller pieces
2018-02-06 20:02:51 +08:00
};
return cvtScaleTab[CV_MAT_DEPTH(ddepth)][CV_MAT_DEPTH(sdepth)];
}
#endif
core: dispatch convert_scale
2019-02-08 22:56:25 +08:00
CV_CPU_OPTIMIZATION_NAMESPACE_END
} // namespace
Reference in New Issue Copy Permalink