Skip to content

O
opencv
Commit 87f3451e authored by Vladislav Vinogradov's avatar Vladislav Vinogradov
Browse files
Options

fixed warnings

parent 7106513b
Hide whitespace changes
Inline Side-by-side
Showing with 81 additions and 44 deletions
modules/gpu/src/cuda/bf_knnmatch.cu
View file @ 87f3451e
...@@ -43,6 +43,7 @@ ...@@ -43,6 +43,7 @@
#include "internal_shared.hpp" #include "internal_shared.hpp"
#include "opencv2/gpu/device/limits.hpp" #include "opencv2/gpu/device/limits.hpp"
#include "opencv2/gpu/device/vec_distance.hpp" #include "opencv2/gpu/device/vec_distance.hpp"
#include "opencv2/gpu/device/datamov_utils.hpp"
using namespace cv::gpu; using namespace cv::gpu;
using namespace cv::gpu::device; using namespace cv::gpu::device;
...@@ -235,7 +236,15 @@ namespace cv { namespace gpu { namespace bf_knnmatch ...@@ -235,7 +236,15 @@ namespace cv { namespace gpu { namespace bf_knnmatch
{ {
const int loadX = threadIdx.x + i * BLOCK_SIZE; const int loadX = threadIdx.x + i * BLOCK_SIZE;
s_train[threadIdx.x * BLOCK_SIZE + threadIdx.y] = loadX < train.cols ? train.ptr(min(t * BLOCK_SIZE + threadIdx.y, train.rows - 1))[loadX] : 0; s_train[threadIdx.x * BLOCK_SIZE + threadIdx.y] = 0;
if (loadX < train.cols)
{
T val;
ForceGlob<T>::Load(train.ptr(min(t * BLOCK_SIZE + threadIdx.y, train.rows - 1)), loadX, val);
s_train[threadIdx.x * BLOCK_SIZE + threadIdx.y] = val;
}
__syncthreads(); __syncthreads();
...@@ -402,15 +411,18 @@ namespace cv { namespace gpu { namespace bf_knnmatch ...@@ -402,15 +411,18 @@ namespace cv { namespace gpu { namespace bf_knnmatch
{ {
const int loadX = threadIdx.x + i * BLOCK_SIZE; const int loadX = threadIdx.x + i * BLOCK_SIZE;
s_query[threadIdx.y * BLOCK_SIZE + threadIdx.x] = 0;
s_train[threadIdx.x * BLOCK_SIZE + threadIdx.y] = 0;
if (loadX < query.cols) if (loadX < query.cols)
{ {
s_query[threadIdx.y * BLOCK_SIZE + threadIdx.x] = query.ptr(min(queryIdx, query.rows - 1))[loadX]; T val;
s_train[threadIdx.x * BLOCK_SIZE + threadIdx.y] = train.ptr(min(t * BLOCK_SIZE + threadIdx.y, train.rows - 1))[loadX];
} ForceGlob<T>::Load(query.ptr(min(queryIdx, query.rows - 1)), loadX, val);
else s_query[threadIdx.y * BLOCK_SIZE + threadIdx.x] = val;
{
s_query[threadIdx.y * BLOCK_SIZE + threadIdx.x] = 0; ForceGlob<T>::Load(train.ptr(min(t * BLOCK_SIZE + threadIdx.y, train.rows - 1)), loadX, val);
s_train[threadIdx.x * BLOCK_SIZE + threadIdx.y] = 0; s_train[threadIdx.x * BLOCK_SIZE + threadIdx.y] = val;
} }
__syncthreads(); __syncthreads();
...@@ -573,15 +585,18 @@ namespace cv { namespace gpu { namespace bf_knnmatch ...@@ -573,15 +585,18 @@ namespace cv { namespace gpu { namespace bf_knnmatch
{ {
const int loadX = threadIdx.x + i * BLOCK_SIZE; const int loadX = threadIdx.x + i * BLOCK_SIZE;
s_query[threadIdx.y * BLOCK_SIZE + threadIdx.x] = 0;
s_train[threadIdx.x * BLOCK_SIZE + threadIdx.y] = 0;
if (loadX < query.cols) if (loadX < query.cols)
{ {
s_query[threadIdx.y * BLOCK_SIZE + threadIdx.x] = query.ptr(min(queryIdx, query.rows - 1))[loadX]; T val;
s_train[threadIdx.x * BLOCK_SIZE + threadIdx.y] = train.ptr(min(t * BLOCK_SIZE + threadIdx.y, train.rows - 1))[loadX];
} ForceGlob<T>::Load(query.ptr(min(queryIdx, query.rows - 1)), loadX, val);
else s_query[threadIdx.y * BLOCK_SIZE + threadIdx.x] = val;
{
s_query[threadIdx.y * BLOCK_SIZE + threadIdx.x] = 0; ForceGlob<T>::Load(train.ptr(min(t * BLOCK_SIZE + threadIdx.y, train.rows - 1)), loadX, val);
s_train[threadIdx.x * BLOCK_SIZE + threadIdx.y] = 0; s_train[threadIdx.x * BLOCK_SIZE + threadIdx.y] = val;
} }
__syncthreads(); __syncthreads();
......
modules/gpu/src/cuda/bf_match.cu
View file @ 87f3451e
...@@ -43,6 +43,7 @@ ...@@ -43,6 +43,7 @@
#include "internal_shared.hpp" #include "internal_shared.hpp"
#include "opencv2/gpu/device/limits.hpp" #include "opencv2/gpu/device/limits.hpp"
#include "opencv2/gpu/device/vec_distance.hpp" #include "opencv2/gpu/device/vec_distance.hpp"
#include "opencv2/gpu/device/datamov_utils.hpp"
using namespace cv::gpu; using namespace cv::gpu;
using namespace cv::gpu::device; using namespace cv::gpu::device;
...@@ -110,7 +111,15 @@ namespace cv { namespace gpu { namespace bf_match ...@@ -110,7 +111,15 @@ namespace cv { namespace gpu { namespace bf_match
{ {
const int loadX = threadIdx.x + i * BLOCK_SIZE; const int loadX = threadIdx.x + i * BLOCK_SIZE;
s_train[threadIdx.x * BLOCK_SIZE + threadIdx.y] = loadX < train.cols ? train.ptr(min(t * BLOCK_SIZE + threadIdx.y, train.rows - 1))[loadX] : 0; s_train[threadIdx.x * BLOCK_SIZE + threadIdx.y] = 0;
if (loadX < train.cols)
{
T val;
ForceGlob<T>::Load(train.ptr(min(t * BLOCK_SIZE + threadIdx.y, train.rows - 1)), loadX, val);
s_train[threadIdx.x * BLOCK_SIZE + threadIdx.y] = val;
}
__syncthreads(); __syncthreads();
...@@ -258,15 +267,18 @@ namespace cv { namespace gpu { namespace bf_match ...@@ -258,15 +267,18 @@ namespace cv { namespace gpu { namespace bf_match
{ {
const int loadX = threadIdx.x + i * BLOCK_SIZE; const int loadX = threadIdx.x + i * BLOCK_SIZE;
s_query[threadIdx.y * BLOCK_SIZE + threadIdx.x] = 0;
s_train[threadIdx.x * BLOCK_SIZE + threadIdx.y] = 0;
if (loadX < query.cols) if (loadX < query.cols)
{ {
s_query[threadIdx.y * BLOCK_SIZE + threadIdx.x] = query.ptr(min(queryIdx, query.rows - 1))[loadX]; T val;
s_train[threadIdx.x * BLOCK_SIZE + threadIdx.y] = train.ptr(min(t * BLOCK_SIZE + threadIdx.y, train.rows - 1))[loadX];
} ForceGlob<T>::Load(query.ptr(min(queryIdx, query.rows - 1)), loadX, val);
else s_query[threadIdx.y * BLOCK_SIZE + threadIdx.x] = val;
{
s_query[threadIdx.y * BLOCK_SIZE + threadIdx.x] = 0; ForceGlob<T>::Load(train.ptr(min(t * BLOCK_SIZE + threadIdx.y, train.rows - 1)), loadX, val);
s_train[threadIdx.x * BLOCK_SIZE + threadIdx.y] = 0; s_train[threadIdx.x * BLOCK_SIZE + threadIdx.y] = val;
} }
__syncthreads(); __syncthreads();
...@@ -410,15 +422,18 @@ namespace cv { namespace gpu { namespace bf_match ...@@ -410,15 +422,18 @@ namespace cv { namespace gpu { namespace bf_match
{ {
const int loadX = threadIdx.x + i * BLOCK_SIZE; const int loadX = threadIdx.x + i * BLOCK_SIZE;
s_query[threadIdx.y * BLOCK_SIZE + threadIdx.x] = 0;
s_train[threadIdx.x * BLOCK_SIZE + threadIdx.y] = 0;
if (loadX < query.cols) if (loadX < query.cols)
{ {
s_query[threadIdx.y * BLOCK_SIZE + threadIdx.x] = query.ptr(min(queryIdx, query.rows - 1))[loadX]; T val;
s_train[threadIdx.x * BLOCK_SIZE + threadIdx.y] = train.ptr(min(t * BLOCK_SIZE + threadIdx.y, train.rows - 1))[loadX];
} ForceGlob<T>::Load(query.ptr(min(queryIdx, query.rows - 1)), loadX, val);
else s_query[threadIdx.y * BLOCK_SIZE + threadIdx.x] = val;
{
s_query[threadIdx.y * BLOCK_SIZE + threadIdx.x] = 0; ForceGlob<T>::Load(train.ptr(min(t * BLOCK_SIZE + threadIdx.y, train.rows - 1)), loadX, val);
s_train[threadIdx.x * BLOCK_SIZE + threadIdx.y] = 0; s_train[threadIdx.x * BLOCK_SIZE + threadIdx.y] = val;
} }
__syncthreads(); __syncthreads();
......
modules/gpu/src/cuda/bf_radius_match.cu
View file @ 87f3451e
...@@ -43,6 +43,7 @@ ...@@ -43,6 +43,7 @@
#include "internal_shared.hpp" #include "internal_shared.hpp"
#include "opencv2/gpu/device/limits.hpp" #include "opencv2/gpu/device/limits.hpp"
#include "opencv2/gpu/device/vec_distance.hpp" #include "opencv2/gpu/device/vec_distance.hpp"
#include "opencv2/gpu/device/datamov_utils.hpp"
using namespace cv::gpu; using namespace cv::gpu;
using namespace cv::gpu::device; using namespace cv::gpu::device;
...@@ -73,15 +74,18 @@ namespace cv { namespace gpu { namespace bf_radius_match ...@@ -73,15 +74,18 @@ namespace cv { namespace gpu { namespace bf_radius_match
{ {
const int loadX = threadIdx.x + i * BLOCK_SIZE; const int loadX = threadIdx.x + i * BLOCK_SIZE;
s_query[threadIdx.y * BLOCK_SIZE + threadIdx.x] = 0;
s_train[threadIdx.x * BLOCK_SIZE + threadIdx.y] = 0;
if (loadX < query.cols) if (loadX < query.cols)
{ {
s_query[threadIdx.y * BLOCK_SIZE + threadIdx.x] = query.ptr(min(queryIdx, query.rows - 1))[loadX]; T val;
s_train[threadIdx.x * BLOCK_SIZE + threadIdx.y] = train.ptr(min(blockIdx.x * BLOCK_SIZE + threadIdx.y, train.rows - 1))[loadX];
} ForceGlob<T>::Load(query.ptr(min(queryIdx, query.rows - 1)), loadX, val);
else s_query[threadIdx.y * BLOCK_SIZE + threadIdx.x] = val;
{
s_query[threadIdx.y * BLOCK_SIZE + threadIdx.x] = 0; ForceGlob<T>::Load(train.ptr(min(blockIdx.x * BLOCK_SIZE + threadIdx.y, train.rows - 1)), loadX, val);
s_train[threadIdx.x * BLOCK_SIZE + threadIdx.y] = 0; s_train[threadIdx.x * BLOCK_SIZE + threadIdx.y] = val;
} }
__syncthreads(); __syncthreads();
...@@ -181,15 +185,18 @@ namespace cv { namespace gpu { namespace bf_radius_match ...@@ -181,15 +185,18 @@ namespace cv { namespace gpu { namespace bf_radius_match
{ {
const int loadX = threadIdx.x + i * BLOCK_SIZE; const int loadX = threadIdx.x + i * BLOCK_SIZE;
s_query[threadIdx.y * BLOCK_SIZE + threadIdx.x] = 0;
s_train[threadIdx.x * BLOCK_SIZE + threadIdx.y] = 0;
if (loadX < query.cols) if (loadX < query.cols)
{ {
s_query[threadIdx.y * BLOCK_SIZE + threadIdx.x] = query.ptr(min(queryIdx, query.rows - 1))[loadX]; T val;
s_train[threadIdx.x * BLOCK_SIZE + threadIdx.y] = train.ptr(min(blockIdx.x * BLOCK_SIZE + threadIdx.y, train.rows - 1))[loadX];
} ForceGlob<T>::Load(query.ptr(min(queryIdx, query.rows - 1)), loadX, val);
else s_query[threadIdx.y * BLOCK_SIZE + threadIdx.x] = val;
{
s_query[threadIdx.y * BLOCK_SIZE + threadIdx.x] = 0; ForceGlob<T>::Load(train.ptr(min(blockIdx.x * BLOCK_SIZE + threadIdx.y, train.rows - 1)), loadX, val);
s_train[threadIdx.x * BLOCK_SIZE + threadIdx.y] = 0; s_train[threadIdx.x * BLOCK_SIZE + threadIdx.y] = val;
} }
__syncthreads(); __syncthreads();
......
Markdown is supported
0% or
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment