Commit 13c20ce5 authored by Gaurav Kukreja's avatar Gaurav Kukreja

Merge branch 'master' of https://github.com/gkernel/cuda_lab

parents 83fb99da 4f91bfa7
<?xml version="1.0" encoding="UTF-8" standalone="no"?>
<?fileVersion 4.0.0?><cproject storage_type_id="org.eclipse.cdt.core.XmlProjectDescriptionStorage">
<storageModule moduleId="org.eclipse.cdt.core.settings">
<cconfiguration id="com.nvidia.cuda.ide.toolchain.base.629067988">
<storageModule buildSystemId="org.eclipse.cdt.managedbuilder.core.configurationDataProvider" id="com.nvidia.cuda.ide.toolchain.base.629067988" moduleId="org.eclipse.cdt.core.settings" name="Default">
<externalSettings/>
<extensions>
<extension id="com.nvidia.cuda.ide.cubin" point="org.eclipse.cdt.core.BinaryParser"/>
<extension id="com.nvidia.cuda.ide.elf" point="org.eclipse.cdt.core.BinaryParser"/>
<extension id="com.nvidia.cuda.ide.macho" point="org.eclipse.cdt.core.BinaryParser"/>
<extension id="nvcc.errorParser" point="org.eclipse.cdt.core.ErrorParser"/>
<extension id="org.eclipse.cdt.core.VCErrorParser" point="org.eclipse.cdt.core.ErrorParser"/>
<extension id="org.eclipse.cdt.core.GmakeErrorParser" point="org.eclipse.cdt.core.ErrorParser"/>
<extension id="org.eclipse.cdt.core.GCCErrorParser" point="org.eclipse.cdt.core.ErrorParser"/>
<extension id="org.eclipse.cdt.core.GASErrorParser" point="org.eclipse.cdt.core.ErrorParser"/>
<extension id="org.eclipse.cdt.core.GLDErrorParser" point="org.eclipse.cdt.core.ErrorParser"/>
</extensions>
</storageModule>
<storageModule moduleId="cdtBuildSystem" version="4.0.0">
<configuration buildProperties="" id="com.nvidia.cuda.ide.toolchain.base.629067988" name="Default" parent="org.eclipse.cdt.build.core.emptycfg">
<folderInfo id="com.nvidia.cuda.ide.toolchain.base.629067988.1442604378" name="/" resourcePath="">
<toolChain id="com.nvidia.cuda.ide.toolchain.base.1103941143" name="com.nvidia.cuda.ide.toolchain.base" superClass="com.nvidia.cuda.ide.toolchain.base">
<targetPlatform archList="all" binaryParser="com.nvidia.cuda.ide.elf;com.nvidia.cuda.ide.macho;com.nvidia.cuda.ide.cubin" id="com.nvidia.cuda.ide.targetPlatform.516467594" isAbstract="false" name="Debug Platform" osList="linux,macosx" superClass="com.nvidia.cuda.ide.targetPlatform"/>
<builder id="com.nvidia.cuda.ide.builder.321422308" managedBuildOn="false" name="CUDA Toolkit 5.5 Builder.Default" superClass="com.nvidia.cuda.ide.builder"/>
<tool id="nvcc.compiler.base.429621792" name="NVCC Compiler" superClass="nvcc.compiler.base">
<option id="nvcc.compiler.pic.817879281" superClass="nvcc.compiler.pic"/>
</tool>
<tool id="nvcc.linker.base.1610352792" name="NVCC Linker" superClass="nvcc.linker.base"/>
<tool id="nvcc.archiver.base.1029091044" name="NVCC Archiver" superClass="nvcc.archiver.base"/>
<tool id="com.nvidia.host.assembler.862361762" name="Host Assembler" superClass="com.nvidia.host.assembler"/>
</toolChain>
</folderInfo>
</configuration>
</storageModule>
<storageModule moduleId="com.nvidia.cuda.ide.build.project.ICudaProjectConfiguration"/>
<storageModule moduleId="org.eclipse.cdt.core.externalSettings"/>
</cconfiguration>
</storageModule>
<storageModule moduleId="cdtBuildSystem" version="4.0.0">
<project id="ex8.null.681844644" name="ex8"/>
</storageModule>
<storageModule moduleId="scannerConfiguration">
<autodiscovery enabled="true" problemReportingEnabled="true" selectedProfileId=""/>
</storageModule>
<storageModule moduleId="org.eclipse.cdt.core.LanguageSettingsProviders"/>
</cproject>
<?xml version="1.0" encoding="UTF-8"?>
<projectDescription>
<name>ex8</name>
<comment></comment>
<projects>
</projects>
<buildSpec>
<buildCommand>
<name>org.eclipse.cdt.managedbuilder.core.genmakebuilder</name>
<triggers>clean,full,incremental,</triggers>
<arguments>
</arguments>
</buildCommand>
<buildCommand>
<name>org.eclipse.cdt.managedbuilder.core.ScannerConfigBuilder</name>
<triggers>full,incremental,</triggers>
<arguments>
</arguments>
</buildCommand>
</buildSpec>
<natures>
<nature>org.eclipse.cdt.core.cnature</nature>
<nature>org.eclipse.cdt.core.ccnature</nature>
<nature>org.eclipse.cdt.managedbuilder.core.managedBuildNature</nature>
<nature>org.eclipse.cdt.managedbuilder.core.ScannerConfigNature</nature>
</natures>
</projectDescription>
all: main
main: main.cu aux.cu aux.h Makefile
nvcc -o main main.cu aux.cu --ptxas-options=-v --use_fast_math --compiler-options -Wall -lopencv_highgui -lopencv_core
// ###
// ###
// ### Practical Course: GPU Programming in Computer Vision
// ###
// ###
// ### Technical University Munich, Computer Vision Group
// ### Winter Semester 2013/2014, March 3 - April 4
// ###
// ###
// ### Evgeny Strekalovskiy, Maria Klodt, Jan Stuehmer, Mohamed Souiai
// ###
// ###
// ###
// ### THIS FILE IS SUPPOSED TO REMAIN UNCHANGED
// ###
// ###
#include "aux.h"
#include <cstdlib>
#include <iostream>
using std::stringstream;
using std::cerr;
using std::cout;
using std::endl;
using std::string;
// parameter processing: template specialization for T=bool
template<>
bool getParam<bool>(std::string param, bool &var, int argc, char **argv)
{
const char *c_param = param.c_str();
for(int i=argc-1; i>=1; i--)
{
if (argv[i][0]!='-') continue;
if (strcmp(argv[i]+1, c_param)==0)
{
if (!(i+1<argc) || argv[i+1][0]=='-') { var = true; return true; }
std::stringstream ss;
ss << argv[i+1];
ss >> var;
return (bool)ss;
}
}
return false;
}
// opencv helpers
void convert_layered_to_interleaved(float *aOut, const float *aIn, int w, int h, int nc)
{
if (nc==1) { memcpy(aOut, aIn, w*h*sizeof(float)); return; }
size_t nOmega = (size_t)w*h;
for (int y=0; y<h; y++)
{
for (int x=0; x<w; x++)
{
for (int c=0; c<nc; c++)
{
aOut[(nc-1-c) + nc*(x + (size_t)w*y)] = aIn[x + (size_t)w*y + nOmega*c];
}
}
}
}
void convert_layered_to_mat(cv::Mat &mOut, const float *aIn)
{
convert_layered_to_interleaved((float*)mOut.data, aIn, mOut.cols, mOut.rows, mOut.channels());
}
void convert_interleaved_to_layered(float *aOut, const float *aIn, int w, int h, int nc)
{
if (nc==1) { memcpy(aOut, aIn, w*h*sizeof(float)); return; }
size_t nOmega = (size_t)w*h;
for (int y=0; y<h; y++)
{
for (int x=0; x<w; x++)
{
for (int c=0; c<nc; c++)
{
aOut[x + (size_t)w*y + nOmega*c] = aIn[(nc-1-c) + nc*(x + (size_t)w*y)];
}
}
}
}
void convert_mat_to_layered(float *aOut, const cv::Mat &mIn)
{
convert_interleaved_to_layered(aOut, (float*)mIn.data, mIn.cols, mIn.rows, mIn.channels());
}
void showImage(string title, const cv::Mat &mat, int x, int y)
{
const char *wTitle = title.c_str();
cv::namedWindow(wTitle, CV_WINDOW_AUTOSIZE);
cvMoveWindow(wTitle, x, y);
cv::imshow(wTitle, mat);
}
// adding Gaussian noise
float noise(float sigma)
{
float x1 = (float)rand()/RAND_MAX;
float x2 = (float)rand()/RAND_MAX;
return sigma * sqrtf(-2*log(std::max(x1,0.000001f)))*cosf(2*M_PI*x2);
}
void addNoise(cv::Mat &m, float sigma)
{
float *data = (float*)m.data;
int w = m.cols;
int h = m.rows;
int nc = m.channels();
size_t n = (size_t)w*h*nc;
for(size_t i=0; i<n; i++)
{
data[i] += noise(sigma);
}
}
// cuda error checking
string prev_file = "";
int prev_line = 0;
void cuda_check(string file, int line)
{
cudaError_t e = cudaGetLastError();
if (e != cudaSuccess)
{
cout << endl << file << ", line " << line << ": " << cudaGetErrorString(e) << " (" << e << ")" << endl;
if (prev_line>0) cout << "Previous CUDA call:" << endl << prev_file << ", line " << prev_line << endl;
exit(1);
}
prev_file = file;
prev_line = line;
}
// ###
// ###
// ### Practical Course: GPU Programming in Computer Vision
// ###
// ###
// ### Technical University Munich, Computer Vision Group
// ### Winter Semester 2013/2014, March 3 - April 4
// ###
// ###
// ### Evgeny Strekalovskiy, Maria Klodt, Jan Stuehmer, Mohamed Souiai
// ###
// ###
// ###
// ### THIS FILE IS SUPPOSED TO REMAIN UNCHANGED
// ###
// ###
#ifndef AUX_H
#define AUX_H
#include <cuda_runtime.h>
#include <ctime>
#include <opencv2/highgui/highgui.hpp>
#include <opencv2/imgproc/imgproc.hpp>
#include <string>
#include <sstream>
// parameter processing
template<typename T>
bool getParam(std::string param, T &var, int argc, char **argv)
{
const char *c_param = param.c_str();
for(int i=argc-1; i>=1; i--)
{
if (argv[i][0]!='-') continue;
if (strcmp(argv[i]+1, c_param)==0)
{
if (!(i+1<argc)) continue;
std::stringstream ss;
ss << argv[i+1];
ss >> var;
return (bool)ss;
}
}
return false;
}
// opencv helpers
void convert_mat_to_layered(float *aOut, const cv::Mat &mIn);
void convert_layered_to_mat(cv::Mat &mOut, const float *aIn);
void showImage(std::string title, const cv::Mat &mat, int x, int y);
// adding Gaussian noise
void addNoise(cv::Mat &m, float sigma);
// measuring time
class Timer
{
public:
Timer() : tStart(0), running(false), sec(0.f)
{
}
void start()
{
tStart = clock();
running = true;
}
void end()
{
if (!running) { sec = 0; return; }
cudaDeviceSynchronize();
clock_t tEnd = clock();
sec = (float)(tEnd - tStart) / CLOCKS_PER_SEC;
running = false;
}
float get()
{
if (running) end();
return sec;
}
private:
clock_t tStart;
bool running;
float sec;
};
// cuda error checking
#define CUDA_CHECK cuda_check(__FILE__,__LINE__)
void cuda_check(std::string file, int line);
#endif // AUX_H
This diff is collapsed.
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