X-Git-Url: https://bilbo.iut-bm.univ-fcomte.fr/and/gitweb/book_gpu.git/blobdiff_plain/4eb0d6980c190aa2e92700dd01c5f685405590bd..c5a3e45c3216da9020c633ff1e344d6176e6ba98:/BookGPU/Chapters/chapter2/ex3.cu

diff --git a/BookGPU/Chapters/chapter2/ex3.cu b/BookGPU/Chapters/chapter2/ex3.cu
index fbdf3a2..cddcc30 100644
--- a/BookGPU/Chapters/chapter2/ex3.cu
+++ b/BookGPU/Chapters/chapter2/ex3.cu
@@ -6,16 +6,12 @@
 #include "cutil_inline.h"
 #include <cublas_v2.h>
 
-
 const int width=16;
 const int nbTh=width*width;
 
 const int size=1024;
 const 	int sizeMat=size*size;
 
-
-
-
 __global__ 
 void matmul(float *d_A, float *d_B, float *d_C) {
 	int i= blockIdx.y*blockDim.y+ threadIdx.y;
@@ -26,15 +22,10 @@ void matmul(float *d_A, float *d_B, float *d_C) {
 		sum+=d_A[i*size+k]*d_B[k*size+j];
 	}	
 	d_C[i*size+j]=sum;
-
 }
 
-
-
-
 int main( int argc, char** argv) 
 {
-
 	float *h_arrayA=(float*)malloc(sizeMat*sizeof(float));
 	float *h_arrayB=(float*)malloc(sizeMat*sizeof(float));
 	float *h_arrayC=(float*)malloc(sizeMat*sizeof(float));
@@ -46,9 +37,7 @@ int main( int argc, char** argv)
 	cudaMalloc((void**)&d_arrayB,sizeMat*sizeof(float));
 	cudaMalloc((void**)&d_arrayC,sizeMat*sizeof(float));
 
-
 	srand48(32);
-
 	for(int i=0;i<sizeMat;i++) {
 		h_arrayA[i]=drand48();
 		h_arrayB[i]=drand48();
@@ -61,7 +50,6 @@ int main( int argc, char** argv)
 	cudaMemcpy(d_arrayB,h_arrayB, sizeMat * sizeof(float), cudaMemcpyHostToDevice);
 	cudaMemcpy(d_arrayC,h_arrayC, sizeMat * sizeof(float), cudaMemcpyHostToDevice);
 
-
 	unsigned int timer_cpu = 0;
 	cutilCheckError(cutCreateTimer(&timer_cpu));
   cutilCheckError(cutStartTimer(timer_cpu));
@@ -77,20 +65,13 @@ int main( int argc, char** argv)
 	printf("CPU processing time : %f (ms) \n", cutGetTimerValue(timer_cpu));
 	cutDeleteTimer(timer_cpu);
 
-
-
-
 	unsigned int timer_gpu = 0;
 	cutilCheckError(cutCreateTimer(&timer_gpu));
   cutilCheckError(cutStartTimer(timer_gpu));
 
-
-
 	dim3 dimGrid(size/width,size/width);
 	dim3 dimBlock(width,width);
 
-	printf("%d %d\n",dimGrid.x,dimBlock.x);
-
 	matmul<<<dimGrid,dimBlock>>>(d_arrayA,d_arrayB,d_arrayC);
 	cudaThreadSynchronize();
 	
@@ -100,12 +81,10 @@ int main( int argc, char** argv)
 	
 	cudaMemcpy(h_arrayCgpu,d_arrayC, sizeMat * sizeof(float), cudaMemcpyDeviceToHost);
 	
-	int good=1;
 	for(int i=0;i<sizeMat;i++)
 		if (fabs(h_arrayC[i]-h_arrayCgpu[i])>1e-4)
 			printf("%f %f\n",h_arrayC[i],h_arrayCgpu[i]);
 	
-
 	cudaFree(d_arrayA);
 	cudaFree(d_arrayB);
 	cudaFree(d_arrayC);
@@ -113,7 +92,5 @@ int main( int argc, char** argv)
 	free(h_arrayB);
 	free(h_arrayC);
 	free(h_arrayCgpu);
-
 	return 0;
-
 }