多核编程和并行运算期末复习(4)

for( int i = 0 ; i < SIZE ; i++ ){ a[i] = 1.0 ; }

cudaMemcpy(dev_a, a, SIZE * sizeof(float), cudaMemcpyHostToDevice ) ; cudaEventCreate( &start ) ; cudaEventCreate( &stop ) ; cudaEventRecord( start, 0 ) ;

sumOfSquares_gpu0<<>>( dev_a, SIZE, dev_result ) ; cudaMemcpy(result, dev_result, BLOCK_NUM * sizeof(float), cudaMemcpyDeviceToHost ) ; double sum = 0.0 ;

for(int i = 0 ; i < BLOCK_NUM ; i++ ){ sum += result[i] ; }

cudaEventRecord( stop, 0 ) ; cudaEventSynchronize( stop ) ;

cudaEventElapsedTime( &elapsedTime, start, stop ) ; printf( \ printf( \ cudaEventDestroy( start ) ; cudaEventDestroy( stop ) ; free( a ) ;

cudaFree( dev_a ) ;

cudaFree( dev_result ) ; return 0 ; }

16