<code c> void add)scalar(float s, float * v, int n) {

float * v_gpu;
cudaMalloc(&v_gpu, n*sizeof(float));
cudaMemcpy(v_gpu, v, n*sizeof(float), cudaMemcopyHostToDevice);
dim3 dimBlock(BLOCKSIZE);
dim3 dimGrid(ceil(n / BLOCKSIZE));
add_scalar)gpu

• IBM 360
• asd
• Stanford: Pipeline is visible (simpler HW). Berekey: Pipelin is invisible (HW ore complex)
• e) row: c, c++ colum: Fortran, Matlab
• f) z: zero, s/n: sign/negative, c:carry, o/v: overflow