cublas, 9600GT, double

Accelerated Computing CUDA CUDA Programming and Performance
1

Is it possible to run cublas with double precision on 9600GT?

If it is impossible, how to change this code to run program in emulation mode compiling program with gcc?

#include <ctype.h>

#include <stdio.h>

#include <stdlib.h> 

#include <string.h>

#include <stddef.h>

#include <stdint.h>

#include "cublas.h" 

void CUBLAS_DGEMM (const char *transa, const char *transb, const int *m,

				   const int *n, const int *k, const double *alpha,

				   const devptr_t *devPtrA, const int *lda,

				   const devptr_t *devPtrB, const int *ldb, const double *beta,

				   const devptr_t *devPtrC, const int *ldc)

{

	cublasInit();

	double *A = (double *)(uintptr_t)(*devPtrA);

	double *B = (double *)(uintptr_t)(*devPtrB);

	double *C = (double *)(uintptr_t)(*devPtrC);

	cublasDgemm (transa[0], transb[0], *m, *n, *k, *alpha, A, *lda,

				 B, *ldb, *beta, C, *ldc);

	cublasFree();

	cublasShutdown();

}
2

The 9600GT is a compute capability 1.1 device. It doesn’t support double precision and none of the double precision CUBLAS routines will work.

3

The problem has been solved by importing cublasemu.dll