Er...
A loop of 1 isn't much of a loop. The whole point of this test is to spawn a
boatload of times to stress the ORTE/OMPI spawn integration.
If you want control to make it a shorter test, how about default to looping a
bagillion times (I don't remember how 1300 was chosen), but if argv[1] != NULL,
do an atoi() on that and use that for the loop count.
How's that?
On Aug 10, 2011, at 6:03 PM, eug...@osl.iu.edu wrote:
> Author: eugene
> Date: 2011-08-10 18:03:36 EDT (Wed, 10 Aug 2011)
> New Revision: 1835
>
> Log:
> Cut iterations to dynamic/loop_spawn.c.
>
>
> Text files modified:
> trunk/ibm/dynamic/loop_spawn.c | 2 +-
>
> 1 files changed, 1 insertions(+), 1 deletions(-)
>
> Modified: trunk/ibm/dynamic/loop_spawn.c
> ==============================================================================
> --- trunk/ibm/dynamic/loop_spawn.c (original)
> +++ trunk/ibm/dynamic/loop_spawn.c 2011-08-10 18:03:36 EDT (Wed, 10 Aug
> 2011)
> @@ -8,7 +8,7 @@
> #include "mpi.h"
>
> #define EXE_TEST "loop_child"
> -#define NUM_TESTS 1300
> +#define NUM_TESTS 1 /* 300 */
>
> int main(int argc, char **argv)
> {
> _______________________________________________
> svn-docs-full mailing list
> svn-docs-f...@open-mpi.org
> http://www.open-mpi.org/mailman/listinfo.cgi/svn-docs-full
--
Jeff Squyres
jsquy...@cisco.com
For corporate legal information go to:
http://www.cisco.com/web/about/doing_business/legal/cri/
