Timothy R. Chavez a écrit :
> Looks like you'll be shaving your head, Gilles :) Attach a picture
> please!
Nope.
You soon see that there is a lot to cut off!
>> - if (sched_setaffinity(0, sizeof(mask), &mask) < 0) {
>> - printf("Thread %d: Can't set affinity: %s\n", tid,
>> strerror(errno));
>
> You'll want to use fprintf(stderr,...) here ??
This code is being removed (-)
>> + if (cpuid == -1) {
>> + printf("Thread %d: Can't set affinity.\n", thread_id);
>
> Ditto.
I did not change what was made initially. I should.
>
> debug(DBG_INFO, "RIP: Here lies the quick brown fox. He was not quick
> enough.")
>
> Second, why are you even using it? As it stands, tt doesn't seem at all
> consistent with how reporting is done by this test case as a whole. If
> you want to introduce the debug() macro to the matrix_mult testcase, you
> might consider introducing it as a separate patch that takes the whole
> testcase into consideration?
>
>> - set_affinity(thread_id);
>> pthread_barrier_wait(&mult_start);
>> while (flags[thread_id] != THREAD_DONE) {
>> pthread_mutex_lock(&t->mutex);
>> @@ -219,6 +232,7 @@ void main_thread(void)
>> nsec_t start, end;
>> long smin = 0, smax = 0, cmin = 0, cmax = 0, delta = 0;
>> float savg, cavg;
>> + int cpuid;
>>
>> if ( stats_container_init(&sdat, ITERATIONS) ||
>> stats_container_init(&shist, HIST_BUCKETS) ||
>> @@ -251,7 +265,15 @@ void main_thread(void)
>> }
>> memset(flags, 0, numcpus);
>>
>> - set_affinity(numcpus-1);
>> + cpuid = set_affinity();
>> + if (cpuid == -1) {
>> + printf("Main thread: Can't set affinity.\n");
>
> Same as above.
>
>> + exit(1);
>> + } else {
>> + if (_dbg_lvl)
>> + printf("Main thread: Affinity set to cpu%d\n", cpuid);
>> + }
>> +
>
> Same as above.
>
>> /* run matrix mult operation sequentially */
>> curdat = &sdat;
-------------------------------------------------------------------------
This SF.Net email is sponsored by the Moblin Your Move Developer's challenge
Build the coolest Linux based applications with Moblin SDK & win great prizes
Grand prize is a trip for two to an Open Source event anywhere in the world
http://moblin-contest.org/redirect.php?banner_id=100&url=/
_______________________________________________
Ltp-list mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/ltp-list
