Hello
It may be similar to https://github.com/open-mpi/hwloc/issues/290 but we
weren't able to find the exact issue unfortunately :/
Setting HWLOC_COMPONENTS=-x86 in the environment would disable that code
path, causing the topology to be possibly not as precise.
Brice
Le 08/07/2019 à 20:43,
Hello Brice,
When I was installing PETSc with --download-mpich on a machine with uname
-a = "SunOS n-gage 5.11 illumos-a22312a201 i86pc i386 i86pc", petsc
configure script hung in a conftest. I have the following stack trace. The
compiler is "Sun C 5.10 SunOS_i386 2009/06/03".
Note MPICH was s
