mirror/opencv
1
0
Fork 0
mirror of https://github.com/opencv/opencv.git synced 2024-11-28 13:10:12 +08:00
Code Issues Actions Packages Projects Releases Wiki Activity

rewrite gpu/device/vec_math.hpp file

Browse Source 
old version isn't compiled with CUDA 5.5
new version doesn't depend on functional.hpp
This commit is contained in:
Vladislav Vinogradov 2013-06-04 13:57:35 +04:00
parent 119b7a298d
commit 918381875a
3 changed files with 854 additions and 261 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

1106
modules/gpu/include/opencv2/gpu/device/vec_math.hpp
View File

File diff suppressed because it is too large Load Diff

4
modules/gpu/src/cuda/ccomponetns.cu
View File

@ -153,7 +153,7 @@ namespace cv { namespace gpu { namespace device
template<typename I> __device__ __forceinline__ bool operator() (const I& a, const I& b) const
{
I d = a - b;
I d = saturate_cast<I>(a - b);
return lo.x <= d.x && d.x <= hi.x &&
lo.y <= d.y && d.y <= hi.y &&
lo.z <= d.z && d.z <= hi.z;
@ -169,7 +169,7 @@ namespace cv { namespace gpu { namespace device
template<typename I> __device__ __forceinline__ bool operator() (const I& a, const I& b) const
{
I d = a - b;
I d = saturate_cast<I>(a - b);
return lo.x <= d.x && d.x <= hi.x &&
lo.y <= d.y && d.y <= hi.y &&
lo.z <= d.z && d.z <= hi.z &&

5
modules/gpu/src/cuda/hough.cu
View File

@ -48,6 +48,7 @@
#include "opencv2/gpu/device/common.hpp"
#include "opencv2/gpu/device/emulation.hpp"
#include "opencv2/gpu/device/vec_math.hpp"
#include "opencv2/gpu/device/functional.hpp"
#include "opencv2/gpu/device/limits.hpp"
#include "opencv2/gpu/device/dynamic_smem.hpp"
@ -811,7 +812,7 @@ namespace cv { namespace gpu { namespace device
const int ind = ::atomicAdd(r_sizes + n, 1);
if (ind < maxSize)
r_table(n, ind) = p - templCenter;
r_table(n, ind) = saturate_cast<short2>(p - templCenter);
}
void buildRTable_gpu(const unsigned int* coordList, const float* thetaList, int pointsCount,
@ -855,7 +856,7 @@ namespace cv { namespace gpu { namespace device
for (int j = 0; j < r_row_size; ++j)
{
short2 c = p - r_row[j];
int2 c = p - r_row[j];
c.x = __float2int_rn(c.x * idp);
c.y = __float2int_rn(c.y * idp);