Skip to content
GitLab
Commit bf4a4d02 authored by Pedro Gonnet's avatar Pedro Gonnet
Browse files

different definition of space_subsize, use division to avoid integer overflows... 

different definition of space_subsize, use division to avoid integer overflows on massive cell sizes.


Former-commit-id: b337e2bd4e2166e8cd1d3f93690cb02aa46e87b4
parent aaaad95e
Hide whitespace changes
Inline Side-by-side
src/scheduler.c
View file @ bf4a4d02
......@@ -26,6 +26,8 @@
#include <string.h>
#include <math.h>
#include <pthread.h>
#include <limits.h>
#include <omp.h>
/* MPI headers. */
#ifdef WITH_MPI
......@@ -190,7 +192,7 @@ void scheduler_splittasks ( struct scheduler *s ) {
if ( ci->split ) {
/* Make a sub? */
if ( scheduler_dosub && ci->count < space_subsize ) {
if ( scheduler_dosub && ci->count < space_subsize/ci->count ) {
/* convert to a self-subtask. */
t->type = task_type_sub;
......@@ -248,7 +250,7 @@ void scheduler_splittasks ( struct scheduler *s ) {
/* Replace by a single sub-task? */
if ( scheduler_dosub &&
( ci->count + cj->count ) * sid_scale[sid] < space_subsize &&
ci->count * sid_scale[sid] < space_subsize/cj->count &&
sid != 0 && sid != 2 && sid != 6 && sid != 8 ) {
/* Make this task a sub task. */
......@@ -625,23 +627,23 @@ void scheduler_reweight ( struct scheduler *s ) {
t->weight += wscale * __builtin_popcount( t->flags ) * t->ci->count * ( sizeof(int)*8 - __builtin_clz( t->ci->count ) );
break;
case task_type_self:
t->weight += 10 * t->ci->count * t->ci->count;
t->weight += 1 * t->ci->count * t->ci->count;
break;
case task_type_pair:
if ( t->ci->nodeID != nodeID || t->cj->nodeID != nodeID )
t->weight += 30 * wscale * t->ci->count * t->cj->count * sid_scale[ t->flags ];
t->weight += 3 * wscale * t->ci->count * t->cj->count * sid_scale[ t->flags ];
else
t->weight += 20 * wscale * t->ci->count * t->cj->count * sid_scale[ t->flags ];
t->weight += 2 * wscale * t->ci->count * t->cj->count * sid_scale[ t->flags ];
break;
case task_type_sub:
if ( t->cj != NULL ) {
if ( t->ci->nodeID != nodeID || t->cj->nodeID != nodeID )
t->weight += 30 * wscale * t->ci->count * t->cj->count * sid_scale[ t->flags ];
t->weight += 3 * wscale * t->ci->count * t->cj->count * sid_scale[ t->flags ];
else
t->weight += 20 * wscale * t->ci->count * t->cj->count * sid_scale[ t->flags ];
t->weight += 2 * wscale * t->ci->count * t->cj->count * sid_scale[ t->flags ];
}
else
t->weight += 10 * wscale * t->ci->count * t->ci->count;
t->weight += 1 * wscale * t->ci->count * t->ci->count;
break;
case task_type_ghost:
if ( t->ci == t->ci->super )
......@@ -651,14 +653,21 @@ void scheduler_reweight ( struct scheduler *s ) {
case task_type_kick2:
t->weight += wscale * t->ci->count;
break;
break;
}
if ( t->type == task_type_send_xv || t->type == task_type_send_rho )
t->weight *= 10;
t->weight = INT_MAX / 8;
if ( t->type == task_type_recv_xv || t->type == task_type_recv_rho )
t->weight *= 2;
t->weight *= 1.41;
}
// message( "weighting tasks took %.3f ms." , (double)( getticks() - tic ) / CPU_TPS * 1000 );
/* int min = tasks[0].weight, max = tasks[0].weight;
for ( k = 1 ; k < nr_tasks ; k++ )
if ( tasks[k].weight < min )
min = tasks[k].weight;
else if ( tasks[k].weight > max )
max = tasks[k].weight;
message( "task weights are in [ %i , %i ]." , min , max ); */
}
......@@ -682,6 +691,7 @@ void scheduler_start ( struct scheduler *s , unsigned int mask ) {
for ( k = nr_tasks - 1 ; k >= 0 ; k-- ) {
t = &tasks[ tid[k] ];
t->wait = 0;
t->rid = -1;
if ( !( (1 << t->type) & mask ) || t->skip )
continue;
for ( j = 0 ; j < t->nr_unlock_tasks ; j++ )
......@@ -691,15 +701,15 @@ void scheduler_start ( struct scheduler *s , unsigned int mask ) {
/* Loop over the tasks and enqueue whoever is ready. */
// tic = getticks();
// #pragma omp parallel for schedule(static) private(t)
for ( k = 0 ; k < nr_tasks ; k++ ) {
for ( k = 0 ; k < nr_tasks ; k++) {
t = &tasks[ tid[k] ];
if ( ( (1 << t->type) & mask ) && !t->skip && t->wait == 0 ) {
if ( t->implicit )
for ( j = 0 ; j < t->nr_unlock_tasks ; j++ )
atomic_dec( &t->unlock_tasks[j]->wait );
else
scheduler_enqueue( s , t );
if ( ( (1 << t->type) & mask ) && !t->skip ) {
if ( t->wait == 0 ) {
scheduler_enqueue( s , t );
pthread_cond_broadcast( &s->sleep_cond );
}
else
break;
}
}
// message( "enqueueing tasks took %.3f ms." , (double)( getticks() - tic ) / CPU_TPS * 1000 );
......@@ -722,7 +732,7 @@ void scheduler_enqueue ( struct scheduler *s , struct task *t ) {
#endif
/* Ignore skipped tasks. */
if ( t->skip )
if ( t->skip || atomic_cas( &t->rid , -1 , 0 ) != -1 )
return;
/* If this is an implicit task, just pretend it's done. */
......@@ -764,7 +774,7 @@ void scheduler_enqueue ( struct scheduler *s , struct task *t ) {
}
// message( "recieving %i parts with tag=%i from %i to %i." ,
// t->ci->count , t->flags , t->ci->nodeID , s->nodeID ); fflush(stdout);
qid = 0;
qid = 1 % s->nr_queues;
#else
error( "SWIFT was not compiled with MPI support." );
#endif
......@@ -971,15 +981,17 @@ struct task *scheduler_gettask ( struct scheduler *s , int qid , struct cell *su
}
/* If we failed, take a short nap. */
#ifndef WITH_MPI
#ifdef WITH_MPI
if ( res == NULL && qid > 1 ) {
#else
if ( res == NULL ) {
#endif
pthread_mutex_lock( &s->sleep_mutex );
if ( s->waiting > 0 )
pthread_cond_wait( &s->sleep_cond , &s->sleep_mutex );
pthread_mutex_unlock( &s->sleep_mutex );
}
#endif
}
/* Start the timer on this task, if we got one. */
......
Supports Markdown
0% or .
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment