generic-library/opencv
1
0
Fork 0
Code Issues Pull Requests Releases Wiki Activity

removed ocl::magnitudeSqr

Browse Source
This commit is contained in:
Ilya Lavrenov
2013-09-24 13:39:07 +04:00
parent 0ad03162df
commit bd36e556a1
5 changed files with 0 additions and 342 deletions
Download Patch File Download Diff File Expand all files Collapse all files

88
modules/ocl/src/arithm.cpp
View File

@@ -84,7 +84,6 @@ namespace cv
extern const char *arithm_bitwise_not;
extern const char *arithm_compare_eq;
extern const char *arithm_compare_ne;
extern const char *arithm_magnitudeSqr;
extern const char *arithm_transpose;
extern const char *arithm_flip;
extern const char *arithm_flip_rc;
@@ -1911,93 +1910,6 @@ void cv::ocl::addWeighted(const oclMat &src1, double alpha, const oclMat &src2,
openCLExecuteKernel(clCxt, &arithm_addWeighted, "addWeighted", globalThreads, localThreads, args, -1, depth);
}
void cv::ocl::magnitudeSqr(const oclMat &src1, const oclMat &src2, oclMat &dst)
{
CV_Assert(src1.type() == src2.type() && src1.size() == src2.size() &&
(src1.depth() == CV_32F ));
dst.create(src1.size(), src1.type());
Context *clCxt = src1.clCxt;
int channels = dst.oclchannels();
int depth = dst.depth();
int vector_lengths[4][7] = {{4, 0, 4, 4, 4, 4, 4},
{4, 0, 4, 4, 4, 4, 4},
{4, 0, 4, 4, 4, 4, 4},
{4, 0, 4, 4, 4, 4, 4}
};
size_t vector_length = vector_lengths[channels - 1][depth];
int offset_cols = (dst.offset / dst.elemSize1()) & (vector_length - 1);
int cols = divUp(dst.cols * channels + offset_cols, vector_length);
size_t localThreads[3] = { 256, 1, 1 };
size_t globalThreads[3] = { cols, dst.rows, 1 };
int dst_step1 = dst.cols * dst.elemSize();
vector<pair<size_t , const void *> > args;
args.push_back( make_pair( sizeof(cl_mem), (void *)&src1.data ));
args.push_back( make_pair( sizeof(cl_int), (void *)&src1.step ));
args.push_back( make_pair( sizeof(cl_int), (void *)&src1.offset));
args.push_back( make_pair( sizeof(cl_mem), (void *)&src2.data ));
args.push_back( make_pair( sizeof(cl_int), (void *)&src2.step ));
args.push_back( make_pair( sizeof(cl_int), (void *)&src2.offset));
args.push_back( make_pair( sizeof(cl_mem), (void *)&dst.data ));
args.push_back( make_pair( sizeof(cl_int), (void *)&dst.step ));
args.push_back( make_pair( sizeof(cl_int), (void *)&dst.offset));
args.push_back( make_pair( sizeof(cl_int), (void *)&src1.rows ));
args.push_back( make_pair( sizeof(cl_int), (void *)&cols ));
args.push_back( make_pair( sizeof(cl_int), (void *)&dst_step1 ));
openCLExecuteKernel(clCxt, &arithm_magnitudeSqr, "magnitudeSqr", globalThreads, localThreads, args, 1, depth);
}
void cv::ocl::magnitudeSqr(const oclMat &src1, oclMat &dst)
{
CV_Assert (src1.depth() == CV_32F );
CV_Assert(src1.size() == dst.size());
dst.create(src1.size(), CV_32FC1);
Context *clCxt = src1.clCxt;
int channels = dst.oclchannels();
int depth = dst.depth();
int vector_lengths[4][7] = {{4, 0, 4, 4, 4, 4, 4},
{4, 0, 4, 4, 4, 4, 4},
{4, 0, 4, 4, 4, 4, 4},
{4, 0, 4, 4, 4, 4, 4}
};
size_t vector_length = vector_lengths[channels - 1][depth];
int offset_cols = (dst.offset / dst.elemSize1()) & (vector_length - 1);
int cols = divUp(dst.cols * channels + offset_cols, vector_length);
size_t localThreads[3] = { 256, 1, 1 };
size_t globalThreads[3] = { cols, dst.rows, 1 };
int dst_step1 = dst.cols * dst.elemSize();
vector<pair<size_t , const void *> > args;
args.push_back( make_pair( sizeof(cl_mem), (void *)&src1.data ));
args.push_back( make_pair( sizeof(cl_int), (void *)&src1.step ));
args.push_back( make_pair( sizeof(cl_int), (void *)&src1.offset));
args.push_back( make_pair( sizeof(cl_mem), (void *)&dst.data ));
args.push_back( make_pair( sizeof(cl_int), (void *)&dst.step ));
args.push_back( make_pair( sizeof(cl_int), (void *)&dst.offset));
args.push_back( make_pair( sizeof(cl_int), (void *)&src1.rows ));
args.push_back( make_pair( sizeof(cl_int), (void *)&cols ));
args.push_back( make_pair( sizeof(cl_int), (void *)&dst_step1 ));
openCLExecuteKernel(clCxt, &arithm_magnitudeSqr, "magnitudeSqr", globalThreads, localThreads, args, 2, depth);
}
static void arithmetic_pow_run(const oclMat &src1, double p, oclMat &dst, string kernelName, const char **kernelString)
{
CV_Assert(src1.cols == dst.cols && src1.rows == dst.rows);

177
modules/ocl/src/opencl/arithm_magnitudeSqr.cl
View File

@@ -1,177 +0,0 @@
/*M///////////////////////////////////////////////////////////////////////////////////////
//
// IMPORTANT: READ BEFORE DOWNLOADING, COPYING, INSTALLING OR USING.
//
// By downloading, copying, installing or using the software you agree to this license.
// If you do not agree to this license, do not download, install,
// copy or use the software.
//
//
// License Agreement
// For Open Source Computer Vision Library
//
// Copyright (C) 2010-2012, Institute Of Software Chinese Academy Of Science, all rights reserved.
// Copyright (C) 2010-2012, Advanced Micro Devices, Inc., all rights reserved.
// Third party copyrights are property of their respective owners.
//
// @Authors
// Jia Haipeng, jiahaipeng95@gmail.com
//
// Redistribution and use in source and binary forms, with or without modification,
// are permitted provided that the following conditions are met:
//
// * Redistribution's of source code must retain the above copyright notice,
// this list of conditions and the following disclaimer.
//
// * Redistribution's in binary form must reproduce the above copyright notice,
// this list of conditions and the following disclaimer in the documentation
// and/or other oclMaterials provided with the distribution.
//
// * The name of the copyright holders may not be used to endorse or promote products
// derived from this software without specific prior written permission.
//
// This software is provided by the copyright holders and contributors as is and
// any express or implied warranties, including, but not limited to, the implied
// warranties of merchantability and fitness for a particular purpose are disclaimed.
// In no event shall the Intel Corporation or contributors be liable for any direct,
// indirect, incidental, special, exemplary, or consequential damages
// (including, but not limited to, procurement of substitute goods or services;
// loss of use, data, or profits; or business interruption) however caused
// and on any theory of liability, whether in contract, strict liability,
// or tort (including negligence or otherwise) arising in any way out of
// the use of this softwareif advised of the possibility of such damage.
//
//M*/
#if defined (DOUBLE_SUPPORT)
#pragma OPENCL EXTENSION cl_khr_fp64:enable
#endif
//////////////////////////////////////////////////////////////////////////////////////////////////////
/////////////////////////////////////////////magnitudeSqr//////////////////////////////////////////////
///////////////////////////////////////////////////////////////////////////////////////////////////////
__kernel void magnitudeSqr_C1_D5 (__global float *src1,int src1_step,int src1_offset,
__global float *src2, int src2_step,int src2_offset,
__global float *dst, int dst_step,int dst_offset,
int rows, int cols,int dst_step1)
{
int x = get_global_id(0);
int y = get_global_id(1);
if (x < cols && y < rows)
{
x = x << 2;
#define dst_align ((dst_offset >> 2) & 3)
int src1_index = mad24(y, src1_step, (x << 2) + src1_offset - (dst_align << 2));
int src2_index = mad24(y, src2_step, (x << 2) + src2_offset - (dst_align << 2));
int dst_start = mad24(y, dst_step, dst_offset);
int dst_end = mad24(y, dst_step, dst_offset + dst_step1);
int dst_index = mad24(y, dst_step, dst_offset + (x << 2) -(dst_align << 2));
int src1_index_fix = src1_index < 0 ? 0 : src1_index;
int src2_index_fix = src2_index < 0 ? 0 : src2_index;
float4 src1_data = vload4(0, (__global float *)((__global char *)src1 + src1_index_fix));
float4 src2_data = vload4(0, (__global float *)((__global char *)src2 + src2_index_fix));
if(src1_index < 0)
{
float4 tmp;
tmp.xyzw = (src1_index == -2) ? src1_data.zwxy:src1_data.yzwx;
src1_data.xyzw = (src1_index == -1) ? src1_data.wxyz:tmp.xyzw;
}
if(src2_index < 0)
{
float4 tmp;
tmp.xyzw = (src2_index == -2) ? src2_data.zwxy:src2_data.yzwx;
src2_data.xyzw = (src2_index == -1) ? src2_data.wxyz:tmp.xyzw;
}
float4 dst_data = *((__global float4 *)((__global char *)dst + dst_index));
float4 tmp_data ;
tmp_data.x = src1_data.x * src1_data.x + src2_data.x * src2_data.x;
tmp_data.y = src1_data.y * src1_data.y + src2_data.y * src2_data.y;
tmp_data.z = src1_data.z * src1_data.z + src2_data.z * src2_data.z;
tmp_data.w = src1_data.w * src1_data.w + src2_data.w * src2_data.w;
dst_data.x = ((dst_index + 0 >= dst_start) && (dst_index + 0 < dst_end)) ? tmp_data.x : dst_data.x;
dst_data.y = ((dst_index + 4 >= dst_start) && (dst_index + 4 < dst_end)) ? tmp_data.y : dst_data.y;
dst_data.z = ((dst_index + 8 >= dst_start) && (dst_index + 8 < dst_end)) ? tmp_data.z : dst_data.z;
dst_data.w = ((dst_index + 12 >= dst_start) && (dst_index + 12 < dst_end)) ? tmp_data.w : dst_data.w;
*((__global float4 *)((__global char *)dst + dst_index)) = dst_data;
}
}
#if defined (DOUBLE_SUPPORT)
__kernel void magnitudeSqr_C2_D5 (__global float *src1,int src1_step,int src1_offset,
__global float *dst, int dst_step,int dst_offset,
int rows, int cols,int dst_step1)
{
int x = get_global_id(0);
int y = get_global_id(1);
if (x < cols && y < rows)
{
x = x << 2;
#define dst_align ((dst_offset >> 2) & 3)
int src1_index = mad24(y, src1_step, (x << 3) + src1_offset - (dst_align << 3));
int dst_start = mad24(y, dst_step, dst_offset);
int dst_end = mad24(y, dst_step, dst_offset + dst_step1);
int dst_index = mad24(y, dst_step, dst_offset + (x << 2) -(dst_align << 2));
int src1_index_fix = src1_index < 0 ? 0 : src1_index;
float8 src1_data = vload8(0, (__global float *)((__global char *)src1 + src1_index_fix));
if(src1_index==-6)
src1_data.s01234567 = src1_data.s67012345;
if(src1_index==-4)
src1_data.s01234567 = src1_data.s45670123;
if(src1_index== -2)
src1_data.s01234567 = src1_data.s23456701;
float4 dst_data = *((__global float4 *)((__global char *)dst + dst_index));
float4 tmp_data ;
tmp_data.x = src1_data.s0 * src1_data.s0 + src1_data.s1 * src1_data.s1;
tmp_data.y = src1_data.s2 * src1_data.s2 + src1_data.s3 * src1_data.s3;
tmp_data.z = src1_data.s4 * src1_data.s4 + src1_data.s5 * src1_data.s5;
tmp_data.w = src1_data.s6 * src1_data.s6 + src1_data.s7 * src1_data.s7;
dst_data.x = ((dst_index + 0 >= dst_start) && (dst_index + 0 < dst_end)) ? tmp_data.x : dst_data.x;
dst_data.y = ((dst_index + 4 >= dst_start) && (dst_index + 4 < dst_end)) ? tmp_data.y : dst_data.y;
dst_data.z = ((dst_index + 8 >= dst_start) && (dst_index + 8 < dst_end)) ? tmp_data.z : dst_data.z;
dst_data.w = ((dst_index + 12 >= dst_start) && (dst_index + 12 < dst_end)) ? tmp_data.w : dst_data.w;
*((__global float4 *)((__global char *)dst + dst_index)) = dst_data;
}
}
#endif