Creating nightly hwloc snapshot SVN tarball was a success.
Snapshot: hwloc 1.0.3a1r2481
Start time: Sun Sep 19 21:02:43 EDT 2010
End time: Sun Sep 19 21:04:31 EDT 2010
Your friendly daemon,
Cyrador
ERROR: Command returned a non-zero exist status (trunk):
make distcheck
Start time: Sun Sep 19 21:01:11 EDT 2010
End time: Sun Sep 19 21:02:43 EDT 2010
===
[... previous lines snipped ...]
CC lstopo-lstopo-color.o
Pavan Balaji, le Sun 19 Sep 2010 14:16:26 -0500, a écrit :
> Thanks Samuel. This patch seems to work. There are still some warnings
> with respect to the redefinition of CPU_SET, etc.
Sure, we can't fix pgcc :)
> but it doesn't seem to be causing any runtime errors.
Good!
Samuel
Thanks Samuel. This patch seems to work. There are still some warnings
with respect to the redefinition of CPU_SET, etc., but it doesn't seem
to be causing any runtime errors.
-- Pavan
On 09/19/2010 09:32 AM, Samuel Thibault wrote:
Could you try the attached patch?
I some sense it adds an
Brice Goglin, le Sun 19 Sep 2010 15:44:54 +0200, a écrit :
> I suggest you report the problem to PGI. I looks like it is entirely
> caused by PGI doing strange stuff with sched.h. I'd like a clear
> explanation from PGI of what they're trying to do
I know what they're trying to do: they try to exp
Could you try the attached patch?
I some sense it adds an odd logic, but it might just work around pgcc's
crazyness.
Samuel
Index: src/topology-linux.c
===
--- src/topology-linux.c(révision 2420)
+++ src/topology-linux.c
Pavan Balaji, le Sun 19 Sep 2010 08:26:03 -0500, a écrit :
> Any news on this?
It seems it went off my mailbox. It's probably a good idea to keep open
bug items for such issues. Unfortunately our PGI trial license expired.
Do you know which macro we could check for to detect when being compiled
Hi Brice,
On 09/19/2010 08:44 AM, Brice Goglin wrote:
I suggest you report the problem to PGI. I looks like it is entirely
caused by PGI doing strange stuff with sched.h. I'd like a clear
explanation from PGI of what they're trying to do and some guarantee
that we won't have to add newer hacks e
Hello Pavan,
I suggest you report the problem to PGI. I looks like it is entirely
caused by PGI doing strange stuff with sched.h. I'd like a clear
explanation from PGI of what they're trying to do and some guarantee
that we won't have to add newer hacks everytime they release a new
compiler. I am
Any news on this?
How are other projects using hwloc dealing with this? This failure is
causing the MPICH2 trunk to fail at "make" time with the PGI compilers.
We have a sandbox configure for hwloc, so if it fails at configure time,
MPICH2 can still proceed, but that's not helping in this cas
10 matches
Mail list logo
