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

used new device layer for cv::gpu::minMax

Browse Source
This commit is contained in:
Vladislav Vinogradov
2013-08-26 11:07:34 +04:00
parent b705e0d886
commit 045a856c24
3 changed files with 62 additions and 246 deletions
Download Patch File Download Diff File Expand all files Collapse all files

249
modules/cudaarithm/src/cuda/minmax.cu
View File

@@ -40,208 +40,77 @@
//
//M*/
#if !defined CUDA_DISABLER
#include "opencv2/opencv_modules.hpp"
#include "opencv2/core/cuda/common.hpp"
#include "opencv2/core/cuda/vec_traits.hpp"
#include "opencv2/core/cuda/vec_math.hpp"
#include "opencv2/core/cuda/functional.hpp"
#include "opencv2/core/cuda/reduce.hpp"
#include "opencv2/core/cuda/emulation.hpp"
#include "opencv2/core/cuda/limits.hpp"
#include "opencv2/core/cuda/utility.hpp"
#ifndef HAVE_OPENCV_CUDEV
using namespace cv::cuda;
using namespace cv::cuda::device;
#error "opencv_cudev is required"
namespace minMax
#else
#include "opencv2/cudaarithm.hpp"
#include "opencv2/cudev.hpp"
using namespace cv::cudev;
namespace
{
__device__ unsigned int blocks_finished = 0;
// To avoid shared bank conflicts we convert each value into value of
// appropriate type (32 bits minimum)
template <typename T> struct MinMaxTypeTraits;
template <> struct MinMaxTypeTraits<uchar> { typedef int best_type; };
template <> struct MinMaxTypeTraits<schar> { typedef int best_type; };
template <> struct MinMaxTypeTraits<ushort> { typedef int best_type; };
template <> struct MinMaxTypeTraits<short> { typedef int best_type; };
template <> struct MinMaxTypeTraits<int> { typedef int best_type; };
template <> struct MinMaxTypeTraits<float> { typedef float best_type; };
template <> struct MinMaxTypeTraits<double> { typedef double best_type; };
template <int BLOCK_SIZE, typename R>
struct GlobalReduce
{
static __device__ void run(R& mymin, R& mymax, R* minval, R* maxval, int tid, int bid, R* sminval, R* smaxval)
{
#if __CUDA_ARCH__ >= 200
if (tid == 0)
{
Emulation::glob::atomicMin(minval, mymin);
Emulation::glob::atomicMax(maxval, mymax);
}
#else
__shared__ bool is_last;
if (tid == 0)
{
minval[bid] = mymin;
maxval[bid] = mymax;
__threadfence();
unsigned int ticket = ::atomicAdd(&blocks_finished, 1);
is_last = (ticket == gridDim.x * gridDim.y - 1);
}
__syncthreads();
if (is_last)
{
int idx = ::min(tid, gridDim.x * gridDim.y - 1);
mymin = minval[idx];
mymax = maxval[idx];
const minimum<R> minOp;
const maximum<R> maxOp;
device::reduce<BLOCK_SIZE>(smem_tuple(sminval, smaxval), thrust::tie(mymin, mymax), tid, thrust::make_tuple(minOp, maxOp));
if (tid == 0)
{
minval[0] = mymin;
maxval[0] = mymax;
blocks_finished = 0;
}
}
#endif
}
};
template <int BLOCK_SIZE, typename T, typename R, class Mask>
__global__ void kernel(const PtrStepSz<T> src, const Mask mask, R* minval, R* maxval, const int twidth, const int theight)
{
__shared__ R sminval[BLOCK_SIZE];
__shared__ R smaxval[BLOCK_SIZE];
const int x0 = blockIdx.x * blockDim.x * twidth + threadIdx.x;
const int y0 = blockIdx.y * blockDim.y * theight + threadIdx.y;
const int tid = threadIdx.y * blockDim.x + threadIdx.x;
const int bid = blockIdx.y * gridDim.x + blockIdx.x;
R mymin = numeric_limits<R>::max();
R mymax = -numeric_limits<R>::max();
const minimum<R> minOp;
const maximum<R> maxOp;
for (int i = 0, y = y0; i < theight && y < src.rows; ++i, y += blockDim.y)
{
const T* ptr = src.ptr(y);
for (int j = 0, x = x0; j < twidth && x < src.cols; ++j, x += blockDim.x)
{
if (mask(y, x))
{
const R srcVal = ptr[x];
mymin = minOp(mymin, srcVal);
mymax = maxOp(mymax, srcVal);
}
}
}
device::reduce<BLOCK_SIZE>(smem_tuple(sminval, smaxval), thrust::tie(mymin, mymax), tid, thrust::make_tuple(minOp, maxOp));
GlobalReduce<BLOCK_SIZE, R>::run(mymin, mymax, minval, maxval, tid, bid, sminval, smaxval);
}
const int threads_x = 32;
const int threads_y = 8;
void getLaunchCfg(int cols, int rows, dim3& block, dim3& grid)
{
block = dim3(threads_x, threads_y);
grid = dim3(divUp(cols, block.x * block.y),
divUp(rows, block.y * block.x));
grid.x = ::min(grid.x, block.x);
grid.y = ::min(grid.y, block.y);
}
void getBufSize(int cols, int rows, int& bufcols, int& bufrows)
{
dim3 block, grid;
getLaunchCfg(cols, rows, block, grid);
bufcols = grid.x * grid.y * sizeof(double);
bufrows = 2;
}
__global__ void setDefaultKernel(int* minval_buf, int* maxval_buf)
{
*minval_buf = numeric_limits<int>::max();
*maxval_buf = numeric_limits<int>::min();
}
__global__ void setDefaultKernel(float* minval_buf, float* maxval_buf)
{
*minval_buf = numeric_limits<float>::max();
*maxval_buf = -numeric_limits<float>::max();
}
__global__ void setDefaultKernel(double* minval_buf, double* maxval_buf)
{
*minval_buf = numeric_limits<double>::max();
*maxval_buf = -numeric_limits<double>::max();
}
template <typename R>
void setDefault(R* minval_buf, R* maxval_buf)
{
setDefaultKernel<<<1, 1>>>(minval_buf, maxval_buf);
}
template <typename T>
void run(const PtrStepSzb src, const PtrStepb mask, double* minval, double* maxval, PtrStepb buf)
void minMaxImpl(const GpuMat& _src, const GpuMat& mask, GpuMat& _buf, double* minVal, double* maxVal)
{
typedef typename MinMaxTypeTraits<T>::best_type R;
typedef typename SelectIf<
TypesEquals<T, double>::value,
double,
typename SelectIf<TypesEquals<T, float>::value, float, int>::type
>::type work_type;
dim3 block, grid;
getLaunchCfg(src.cols, src.rows, block, grid);
GpuMat_<T> src(_src);
GpuMat_<work_type> buf(_buf);
const int twidth = divUp(divUp(src.cols, grid.x), block.x);
const int theight = divUp(divUp(src.rows, grid.y), block.y);
R* minval_buf = (R*) buf.ptr(0);
R* maxval_buf = (R*) buf.ptr(1);
setDefault(minval_buf, maxval_buf);
if (mask.data)
kernel<threads_x * threads_y><<<grid, block>>>((PtrStepSz<T>) src, SingleMask(mask), minval_buf, maxval_buf, twidth, theight);
if (mask.empty())
gridFindMinMaxVal(src, buf);
else
kernel<threads_x * threads_y><<<grid, block>>>((PtrStepSz<T>) src, WithOutMask(), minval_buf, maxval_buf, twidth, theight);
gridFindMinMaxVal(src, buf, globPtr<uchar>(mask));
cudaSafeCall( cudaGetLastError() );
work_type data[2];
buf.download(cv::Mat(1, 2, buf.type(), data));
cudaSafeCall( cudaDeviceSynchronize() );
if (minVal)
*minVal = data[0];
R minval_, maxval_;
cudaSafeCall( cudaMemcpy(&minval_, minval_buf, sizeof(R), cudaMemcpyDeviceToHost) );
cudaSafeCall( cudaMemcpy(&maxval_, maxval_buf, sizeof(R), cudaMemcpyDeviceToHost) );
*minval = minval_;
*maxval = maxval_;
if (maxVal)
*maxVal = data[1];
}
template void run<uchar >(const PtrStepSzb src, const PtrStepb mask, double* minval, double* maxval, PtrStepb buf);
template void run<schar >(const PtrStepSzb src, const PtrStepb mask, double* minval, double* maxval, PtrStepb buf);
template void run<ushort>(const PtrStepSzb src, const PtrStepb mask, double* minval, double* maxval, PtrStepb buf);
template void run<short >(const PtrStepSzb src, const PtrStepb mask, double* minval, double* maxval, PtrStepb buf);
template void run<int >(const PtrStepSzb src, const PtrStepb mask, double* minval, double* maxval, PtrStepb buf);
template void run<float >(const PtrStepSzb src, const PtrStepb mask, double* minval, double* maxval, PtrStepb buf);
template void run<double>(const PtrStepSzb src, const PtrStepb mask, double* minval, double* maxval, PtrStepb buf);
}
#endif // CUDA_DISABLER
void cv::cuda::minMax(InputArray _src, double* minVal, double* maxVal, InputArray _mask, GpuMat& buf)
{
typedef void (*func_t)(const GpuMat& _src, const GpuMat& mask, GpuMat& _buf, double* minVal, double* maxVal);
static const func_t funcs[] =
{
minMaxImpl<uchar>,
minMaxImpl<schar>,
minMaxImpl<ushort>,
minMaxImpl<short>,
minMaxImpl<int>,
minMaxImpl<float>,
minMaxImpl<double>
};
GpuMat src = _src.getGpuMat();
GpuMat mask = _mask.getGpuMat();
CV_Assert( src.channels() == 1 );
CV_DbgAssert( mask.empty() || (mask.size() == src.size() && mask.type() == CV_8U) );
const int depth = src.depth();
const int work_type = depth == CV_64F ? CV_64F : depth == CV_32F ? CV_32F : CV_32S;
ensureSizeIsEnough(1, 2, work_type, buf);
const func_t func = funcs[src.depth()];
func(src, mask, buf, minVal, maxVal);
}
#endif