Jack, It might be still an idea for a test to do the workaround suggested in Alan’s doc. Would give more of an idea whether this is a “new new” issue or an “old new” issue (a new facet of a known issue).
Best wishes Joachim On 26 Jan 2017, at 21:25, Joachim Hein <[email protected]<mailto:[email protected]>> wrote: Hi Alan, Jack is using Intel 15.x. The intel doc you quote is specific in stating 16.0.3 and 16.0.4. Joachim On 26 Jan 2017, at 21:20, Alan O'Cais <[email protected]<mailto:[email protected]>> wrote: Yup, known issue https://software.intel.com/en-us/articles/intel-compiler-version-16-not-compatible-with-recent-libcso6 On 26 Jan 2017 9:07 pm, "Jack Perdue" <[email protected]<mailto:[email protected]>> wrote: Howdy all, Has anyone else been experiences problems (e.g. seg faults) after upgrading RHEL/CentOS 7.2 to 7.3? We've been having issues. As has already been reported on this list, the new glibc has issues with Intel 2016 compilers. However I haven't seen a similar report for the 2015 ones I've got a toolchain here that is defined as follows: GCCcore/4.9.4 binutils/2.25-GCCcore-4.9.4 icc/2015.7.235-GCC-4.9.4-2.25 ifort/2015.7.235-GCC-4.9.4-2.25 iccifort/2015.7.235-GCC-4.9.4-2.25 impi/5.1.3.223-iccifort-2015.7.235-GCC-4.9.4-2.25 iimpi/2015F imkl/11.3.4.258-iimpi-2015F intel/2015F I also have a Python-2.7.12-intel-2015F.eb that build fine on RHEL7.2. However it seg faults as soon as it tries to run (self-test) itself. If I downgrade to the 7.2 glibc, then it will build. I've tried rebuilding the entire stack agains the 7.3 glibc and had no problem..... until I got to Python again. And, again, it will build with 7.2's glibc (using dependencies that were all built against 7.3), but it seg faults on the self-test. I/we were just curious if others in HPC were having similar problems. Another issue we've seen is irqbalance issue a LOT of spurious messages along the lines of: Jan 17 13:28:13 tlogin-0501 /usr/sbin/irqbalance: irq 237 affinity_hint and banned cpus confict In perhaps unrelated matters, we've had a number of problems with some locally defined toolchains using the Intel 2017 compilers (which are reported to work). Not sure what's up with that yet waiting to see EB's "sanctioned" toolchain to see if the issues persist. Anyway, just wondering if anyone else was seeing issues with RHEL7.3. -- Jack Perdue Lead Systems Administrator High Performance Research Computing TAMU Division of Research [email protected]<mailto:[email protected]> http://hprc.tamu.edu<http://hprc.tamu.edu/> HPRC Helpdesk: [email protected]<mailto:[email protected]> ------------------------------------------------------------------------------------------------ ------------------------------------------------------------------------------------------------ Forschungszentrum Juelich GmbH 52425 Juelich Sitz der Gesellschaft: Juelich Eingetragen im Handelsregister des Amtsgerichts Dueren Nr. HR B 3498 Vorsitzender des Aufsichtsrats: MinDir Dr. Karl Eugen Huthmacher Geschaeftsfuehrung: Prof. Dr.-Ing. Wolfgang Marquardt (Vorsitzender), Karsten Beneke (stellv. Vorsitzender), Prof. Dr.-Ing. Harald Bolt, Prof. Dr. Sebastian M. Schmidt ------------------------------------------------------------------------------------------------ ------------------------------------------------------------------------------------------------
