Parallel Computing for Data Science

172 CHAPTER 6. SHARED-MEMORY: GPUS

hm = ( int ∗ ) malloc ( msize ) ;

// as a t e s t , f i l l matrix with random 1 s and 0 s

int i , j ;

for ( i = 0 ; i < n ; i ++) {

hm[ n∗ i+i ] = 0 ;

for ( j = 0 ; j < n ; j ++) {

i f ( j != i ) hm[ i ∗n+j ] = rand ( ) % 2 ;

}

}

// more o f the u s ual i n i t i a l i z a t i o n s

cudaMalloc ( ( void ∗∗ )&dm, msize ) ;

// copy host matrix to d e v i c e matrix

cudaMemcpy (dm, hm, msize , cudaMemcpyHostToDevice ) ;

htot = 0 ;

// s e t up d e v i c e t o t a l and i n i t i a l i z e i t

cudaMalloc ( ( void ∗∗ )&dtot , s i ze o f ( int ) ) ;

cudaMemcpy ( dtot ,&htot , s i ze o f ( int ) ,

cudaMemcpyHostToDevice ) ;

// OK,