[hwloc-devel] Create success (hwloc r1.1a1r2280)

2010-06-30 Thread MPI Team
Creating nightly hwloc snapshot SVN tarball was a success. Snapshot: hwloc 1.1a1r2280 Start time: Wed Jun 30 21:01:04 EDT 2010 End time: Wed Jun 30 21:02:57 EDT 2010 Your friendly daemon, Cyrador

Re: [hwloc-devel] standalone-like build inside other projects

2010-06-30 Thread Jeff Squyres
On Jun 30, 2010, at 5:53 AM, Brice Goglin wrote: > Modified: mvapich2/trunk/src/mpid/ch3/channels/mrail/src/hwloc/Makefile.am > === > --- mvapich2/trunk/src/mpid/ch3/channels/mrail/src/hwloc/Makefile.am > 2010-06-29 22:01:58 UTC (

[hwloc-devel] standalone-like build inside other projects

2010-06-30 Thread Brice Goglin
I found this in the mvapich commits: (http://mail.cse.ohio-state.edu/pipermail/mvapich-commit/2010-June/001283.html) No need to build the doc directory Modified: mvapich2/trunk/src/mpid/ch3/channels/mrail/src/hwloc/Makefile.am === --

Re: [hwloc-devel] [hwloc-svn] svn:hwloc r2270

2010-06-30 Thread Samuel Thibault
Brice Goglin, le Wed 30 Jun 2010 09:47:34 +0200, a écrit : > I reverted a commit that was slightly improving Power7 support with old > kernels > but it was also preventing sockets with unknown id from appearing with recent > kernels. The P7 topology is properly detected for real now. With old kern

Re: [hwloc-devel] [hwloc-svn] svn:hwloc r2270

2010-06-30 Thread Brice Goglin
I reverted a commit that was slightly improving Power7 support with old kernels but it was also preventing sockets with unknown id from appearing with recent kernels. The P7 topology is properly detected for real now. With old kernels, the topology will be slightly more broken, but it will never be

Re: [hwloc-devel] [hwloc-svn] svn:hwloc r2270

2010-06-30 Thread Samuel Thibault
Brad Benton, le Tue 29 Jun 2010 17:13:05 -0500, a écrit : > Sure...the tarball and output files are attached.  Ok, so the physical index is explicitly -1, Linux itself doesn't know it. Samuel