OpenCL Performance Evaluation on Modern Multicore CPUs

Algorithm 2

Code snippet of simple application.

/*First Kernel: Vector Addition.*/

#pragma omp parallel for shared(, , ) private ()

for (int  ; < MAX_INDEX; ++){

  = + ;

/*(a) Aligned Second Kernel: Vector Multiplication.*/

#pragma omp parallel for shared(, , ) private ()

for (int  ; < MAX_INDEX; ++){

  = + ;

/*(b) Misaligned Second Kernel: Vector Multiplication.*/

#pragma omp parallel for shared(, , ) private ()

for (int  ; < MAX_INDEX; ++){

  int   = ;

  = + ;