Overnight my (relatively) slow ARM and MIPS testers completed. Most of those have been moved to real hardware (vs QEMU) and I've been able to increase coverage as a result.
I will note that I was not able to test IA64 (yet?) because that system is down. I've emailed the owners of that system and am hopeful that I can test it in the next few days. The only clear Open MPI issues I encountered are the ROMIO-on-OpenBSD and libtool recognition of Solaris Studio 12.5 beta. Each was reported already on this list, and I believe Gilles is now handling both issues. I did encounter one issue that still needs more investigating before I can attribute it to Solaris or to Open MPI: Since I last built Open MPI there x86-64/Solaris systems were updated from Solaris 11.2 to 11.3. I can build this release tarball using "gmake" fine, but "gmake -j4" is failing. It appears to be libtool failing to create a ".libs" directory. My current guess is some issue between the updated Solaris and the NFS server for the filesystem where I am building. -Paul -- Paul H. Hargrove phhargr...@lbl.gov Computer Languages & Systems Software (CLaSS) Group Computer Science Department Tel: +1-510-495-2352 Lawrence Berkeley National Laboratory Fax: +1-510-486-6900