gpu_utils.h File Reference

#include <cuda_runtime.h>
#include <vector>
#include <cstdarg>
#include "cufft.h"
#include "LsqQuadraticFit.h"

Go to the source code of this file.

Classes
class	device_vec< T >
struct	MeasureTime
class	pinned_array< T, flags >

Macros
#define	CUDA_SUPPORTED_FUNC   __device__ __host__
#define	CUBOTH   __device__ __host__
	Kernel is callable from both device and host code. More...

Functions
void	outputTotalGPUMemUse (std::string info="")
void	CheckCUDAError (cufftResult_t err)
void	CheckCUDAError (cudaError_t err)
void	CheckCUDAError ()
void	dbgCUDAErrorCheck (cudaError_t e)
void	DbgCopyResult (device_vec< float2 > src, std::vector< std::complex< float > > &dst)
void	DbgCopyResult (device_vec< float > src, std::vector< float > &dst)
void	DbgOutputVectorToFile (std::string loc, device_vec< float > &src, bool append)

Macro Definition Documentation

CUBOTH

#define CUBOTH   __device__ __host__

Kernel is callable from both device and host code.

Definition at line 16 of file gpu_utils.h.

CUDA_SUPPORTED_FUNC

#define CUDA_SUPPORTED_FUNC   __device__ __host__

Definition at line 12 of file gpu_utils.h.

Function Documentation

CheckCUDAError() [1/3]

void CheckCUDAError ( cufftResult_t  err )

inline

Definition at line 36 of file gpu_utils.h.

{
    if (err != CUFFT_SUCCESS) {
        outputTotalGPUMemUse("CUFFT Error");
        throw std::runtime_error(SPrintf("CUDA error: CUFFT failed (%d)\n",err));
    }
}

CheckCUDAError() [2/3]

void CheckCUDAError ( cudaError_t  err )

inline

Definition at line 44 of file gpu_utils.h.

{
    if (err != cudaSuccess) {
        const char* errstr = cudaGetErrorString(err);
        throw std::runtime_error(SPrintf("CUDA error: %s\n" ,errstr).c_str());
    }
}

CheckCUDAError() [3/3]

void CheckCUDAError ( )

inline

Definition at line 52 of file gpu_utils.h.

{
    cudaError_t err = cudaGetLastError();
    if (err != cudaSuccess) {
        const char* errstr = cudaGetErrorString(err);
        dbgprintf("CUDA error: %s\n" ,errstr);
    }
}

DbgCopyResult() [1/2]

void DbgCopyResult ( device_vec< float2 >  src, std::vector< std::complex< float > > &  dst  )

inline

Definition at line 268 of file gpu_utils.h.

{} 

DbgCopyResult() [2/2]

void DbgCopyResult ( device_vec< float >  src, std::vector< float > &  dst  )

inline

Definition at line 269 of file gpu_utils.h.

{}

dbgCUDAErrorCheck()

void dbgCUDAErrorCheck ( cudaError_t  e )

inline

Definition at line 63 of file gpu_utils.h.

{}

DbgOutputVectorToFile()

void DbgOutputVectorToFile ( std::string  loc, device_vec< float > &  src, bool  append  )

inline

Definition at line 270 of file gpu_utils.h.

{}

outputTotalGPUMemUse()

void outputTotalGPUMemUse ( std::string  info = "" )

inline

Definition at line 18 of file gpu_utils.h.

{
    // show total memory usage of GPU
    size_t free_byte;
    size_t total_byte;
    cudaError_t cuda_status = cudaMemGetInfo( &free_byte, &total_byte );
    if ( cudaSuccess != cuda_status ){
        dbgprintf("Error: cudaMemGetInfo fails, %s \n", cudaGetErrorString(cuda_status) );
        exit(1);
    }
    double free_db = (double)free_byte;
    double total_db = (double)total_byte;
    double used_db = total_db - free_db;
    dbgprintf("%sused = %2.2f MB, free = %2.2f MB, total = %2.2f MB\n",
        info != "" ? (info+": ").c_str() : "",
        used_db/1024.0/1024.0, free_db/1024.0/1024.0, total_db/1024.0/1024.0);
}