Unknown error code 10100

Accelerated Computing CUDA CUDA Programming and Performance
1

Hi all. Rather new to the CUDA programming, but see it’s popularity and potential.

My problem seems simple, but searching for solutions has turned difficult.

Basically, when I do a cudaHostAlloc() call, the return code is not cudaSuccess, but 10100, which resolves to an unspecified driver error. This is happening at the very beginning of my program, so I don’t think anything else is conflicting at this point. Can anyone point me to what this error code might mean?

Thanks

Zack

2

Hi all. Rather new to the CUDA programming, but see it’s popularity and potential.

My problem seems simple, but searching for solutions has turned difficult.

Basically, when I do a cudaHostAlloc() call, the return code is not cudaSuccess, but 10100, which resolves to an unspecified driver error. This is happening at the very beginning of my program, so I don’t think anything else is conflicting at this point. Can anyone point me to what this error code might mean?

Thanks

Zack

3

Also, here’s the code I’m using. Is it perhaps a hardware problem? This is a multi-user machine at a remote facility, so I can’t just reboot it.

Thanks.

Zack

[codebox]

#include <stdio.h>

#include <cuda_runtime_api.h>

static void HandleError( cudaError_t err, const char* file, int line ){

if( err != cudaSuccess ){

    printf( "%s in %s at line %d\n", cudaGetErrorString(err), file, line );

    exit( EXIT_FAILURE );

}

}

#define HANDLE_ERROR(err) (HandleError(err, FILE, LINE ))

int main( int argc, char **argv ){

int nsamples = 128;

int ndimensions = 3;

float *x;

int err;

err = cudaHostAlloc( (void**)&x, sizeof(float)*nsamples*ndimensions, cudaHostAllocDefault );

printf( "First cudaHostAlloc errno: %d\n", err );

HANDLE_ERROR(cudaHostAlloc( (void**)&x, sizeof(float)*nsamples*ndimensions, cudaHostAllocDefault ));

cudaFreeHost(x);

return 1;

}

[/codebox]

4

Also, here’s the code I’m using. Is it perhaps a hardware problem? This is a multi-user machine at a remote facility, so I can’t just reboot it.

Thanks.

Zack

[codebox]

#include <stdio.h>

#include <cuda_runtime_api.h>

static void HandleError( cudaError_t err, const char* file, int line ){

if( err != cudaSuccess ){

    printf( "%s in %s at line %d\n", cudaGetErrorString(err), file, line );

    exit( EXIT_FAILURE );

}

}

#define HANDLE_ERROR(err) (HandleError(err, FILE, LINE ))

int main( int argc, char **argv ){

int nsamples = 128;

int ndimensions = 3;

float *x;

int err;

err = cudaHostAlloc( (void**)&x, sizeof(float)*nsamples*ndimensions, cudaHostAllocDefault );

printf( "First cudaHostAlloc errno: %d\n", err );

HANDLE_ERROR(cudaHostAlloc( (void**)&x, sizeof(float)*nsamples*ndimensions, cudaHostAllocDefault ));

cudaFreeHost(x);

return 1;

}

[/codebox]

5

Well, the admins report that the driver was running in emulation mode. They’ve tried uninstalling and reinstalling the driver and environment, but are still getting errors. Unfortunately, it’s Friday, so a complete resolution will have to wait until next week.

6

Well, the admins report that the driver was running in emulation mode. They’ve tried uninstalling and reinstalling the driver and environment, but are still getting errors. Unfortunately, it’s Friday, so a complete resolution will have to wait until next week.