THMMY.gr

Time to compute the hash codes            : 0.008600s
Time to compute the morton encoding       : 0.011531s
Time for the truncated radix sort         : 0.027455s
Time to rearrange the particles in memory : 0.018445s

struct timespec requestStart, requestEnd;

clock_gettime(CLOCK_MONOTONIC, &requestStart);
  
// function here

clock_gettime(CLOCK_MONOTONIC, &requestEnd);

double t = ( requestEnd.tv_sec - requestStart.tv_sec )+ ( requestEnd.tv_nsec - requestStart.tv_nsec )/1e9;
    

==6863== 684 errors in context 9 of 20:
==6863== Invalid write of size 4
==6863==    at 0x401775: quantize (in /home/panagiotis/Desktop/pthread-octree-code-0.04/test_octree)
==6863==    by 0x4E40709: start_thread (pthread_create.c:333)
==6863==  Address 0x7f2a040 is 0 bytes after a block of size 12,582,912 alloc'd
==6863==    at 0x4C2DBB6: malloc (vg_replace_malloc.c:299)
==6863==    by 0x400C74: main (in /home/panagiotis/Desktop/pthread-octree-code-0.04/test_octree)
==6863== 

     args[i].morton_codes= morton_codes;
      args[i].sorted_morton_codes= sorted_morton_codes;
      args[i].permutation_vector=permutation_vector;
      args[i].index = index;
      args[i].level_record = level_record;
      args[i].size = size;
      args[i].population_threshold = population_threshold;
      args[i].sft = sft - 3;
      args[i].lv= lv+1;
      args[i].offset = offset;
      pthread_create(&threads[i], NULL, truncated_radix_sort,(void *)&args[i]);  
 

memcpy(permutation_vector, index, (N)*sizeof(unsigned int)); // Copy the pernutation vector
memcpy(sorted_morton_codes, morton_codes, (N)*sizeof(unsigned long int)); // Copy the Morton codes 

struct radix_args *one_args = (struct radix_args *) arg;
   unsigned long int *morton_codes = one_args->morton_codes;
   unsigned long int *sorted_morton_codes = one_args->sorted_morton_codes;
   unsigned int *permutation_vector = one_args->permutation_vector;
   unsigned int *level_record = one_args->level_record;
   unsigned int *index = one_args->index;
   int N = one_args->N;
   int population_threshold = one_args->population_threshold;
   int sft =one_args->sft;
   int lv = one_args->lv;

GCC=icc -lcilkrts -ldl -O3 -std=gnu99

 unsigned long int *morton_codes = my_data->morton_codes;

dereferencing pointer to incomplete type ‘struct thread_data’
     unsigned long int *morton_codes = my_data->morton_codes;
                                              ^

warning #266: function "__cilkrts_init" declared implicitly
    __cilkrts_init();
    ^

__cilkrts_end_cilk();
  __cilkrts_set_param("nworkers",str);
  __cilkrts_init();

 for(int i=0; i<MAXBINS; i++){
      int offset = (i>0) ? BinSizes[i-1] : 0;
      int size = BinSizes[i] - offset;
      //printf("LABEl next labels \n");
      //struct RTD next_RTD;
      next_RTD[i].morton_codes=&my_data->morton_codes[offset];
      next_RTD[i].sorted_morton_codes=&my_data->sorted_morton_codes[offset];
      next_RTD[i].permutation_vector=&my_data->permutation_vector[offset];
      next_RTD[i].index=&my_data->index[offset];
      next_RTD[i].level_record=&my_data->level_record[offset];
      next_RTD[i].N=size;
      next_RTD[i].population_threshold=my_data->population_threshold;
      next_RTD[i].sft=my_data->sft-3;
      next_RTD[i].lv=my_data->lv+1;
	  next_RTD[i].spawn_lvl=my_data->spawn_lvl;
    }