Re: [hwloc-devel] Fwd: BGQ empty topology with MPI

Daniel Ibanez, le Wed 28 Mar 2012 20:23:18 +0200, a écrit : > The machine is back in working order. > I tried this patch and it works great: I get cpus and my whole program runs as > expected. Maybe it could be a safe fallback that we should keep anyway, even if we work out the issue with sysfs.

Re: [hwloc-devel] Fwd: BGQ empty topology with MPI

The machine is back in working order. I tried this patch and it works great: I get cpus and my whole program runs as expected. I'm now looking into what failed in look_sysfscpu. On Sun, Mar 25, 2012 at 2:43 AM, Brice Goglin wrote: > Le 24/03/2012 23:04, Daniel Ibanez a

Re: [hwloc-devel] Fwd: BGQ empty topology with MPI

-BEGIN PGP SIGNED MESSAGE- Hash: SHA1 On 26/03/12 17:14, Brice Goglin wrote: > Thanks, that would explain such a strange behavior. Not a problem. > For the record, you can run "lstopo -v" or even "lstopo -.xml" to > get more info, especially machine attributes. OK, please find

Re: [hwloc-devel] Fwd: BGQ empty topology with MPI

Le 26/03/2012 05:16, Christopher Samuel a écrit : > On 25/03/12 09:04, Daniel Ibanez wrote: > > > Additional printfs confirm that with MPI in the code, > > hwloc_accessat succeeds on the various /sys/ directories, but the > > overall procedure for getting PUs from these fails. Without MPI, > >

Re: [hwloc-devel] Fwd: BGQ empty topology with MPI

-BEGIN PGP SIGNED MESSAGE- Hash: SHA1 On 25/03/12 17:43, Brice Goglin wrote: > But it'd be good to understand what's going on in /sys on this > machine. And I still don't understand why MPI changes things here. My guess (looking at the BG/P CNK kernel code) is that /sys is not present

Re: [hwloc-devel] Fwd: BGQ empty topology with MPI

-BEGIN PGP SIGNED MESSAGE- Hash: SHA1 On 25/03/12 09:04, Daniel Ibanez wrote: > Additional printfs confirm that with MPI in the code, > hwloc_accessat succeeds on the various /sys/ directories, but the > overall procedure for getting PUs from these fails. Without MPI, > access to /sys/

Re: [hwloc-devel] Fwd: BGQ empty topology with MPI

Le 24/03/2012 23:04, Daniel Ibanez a écrit : > The fundamental difference is in > > src/topology-linux.c:3251 > > when this if statement is true, hwloc_setup_pu_level > finds the PU objects. > When it is false, it fails with empty topology. > > I checked HWLOC_LINUX_USE_CPUINFO, > and it is not

Re: [hwloc-devel] Fwd: BGQ empty topology with MPI

The fundamental difference is in src/topology-linux.c:3251 when this if statement is true, hwloc_setup_pu_level finds the PU objects. When it is false, it fails with empty topology. I checked HWLOC_LINUX_USE_CPUINFO, and it is not detected even when I set it from the front end. That means the

Re: [hwloc-devel] Fwd: BGQ empty topology with MPI

Le 22/03/2012 23:33, Daniel Ibanez a écrit : > I've run this test before (didnt keep the results but can run it again). > I got debug output and compared it with the output from a hwloc test > executable > and I noticed that my program did not show any PU objects were discovered. > In my program

Re: [hwloc-devel] Fwd: BGQ empty topology with MPI

I've run this test before (didnt keep the results but can run it again). I got debug output and compared it with the output from a hwloc test executable and I noticed that my program did not show any PU objects were discovered. In my program the first discovered topology is just a Machine object,

Re: [hwloc-devel] Fwd: BGQ empty topology with MPI

Le 22/03/2012 23:07, Daniel Ibanez a écrit : > > I suspected this might be the reason, so I called "nm" > with the static versions of the libraries their compiler wrappers > link against and I could not find the term "hwloc" in the output. > Is this a valid test? If your hwloc is still compiled