[petsc-dev] metis config error on XT4

[Tim Gallagher]

Without knowing how your system is configured, this may not help. But on the Cray systems we use (they have the modules package installed), if we are using a compiler other than GNU (Intel, PGI, etc) but we want to use cmake, we still have to load the gcc module (not the programming environment, just the module). 

On our machines, it means we may load the PrgEnv-Intel module (so cc points to icc, ftn points to ifort, etc) but we also have to load gcc/<version> to get cmake to work. 

Tim 

----- Original Message -----

From: "Aron Ahmadia" <aron.ahmadia at kaust.edu.sa> 
To: "For users of the development version of PETSc" <petsc-dev at mcs.anl.gov> 
Sent: Wednesday, November 30, 2011 3:49:25 PM 
Subject: Re: [petsc-dev] metis config error on XT4 


looks like a broken cmake... 


What happens when you just run: cmake by itself? 


A 


On Wed, Nov 30, 2011 at 11:47 PM, Mark F. Adams < mark.adams at columbia.edu > wrote: 


I'm getting this error on Franklin at NERSC: 
******************************************************************************* 
UNABLE to CONFIGURE with GIVEN OPTIONS (see configure.log for details): 
------------------------------------------------------------------------------- 
Error running configure on METIS: Could not execute "cd /global/u2/m/madams/petsc-dev/externalpackages/metis-5.0.2 && make distclean && make config prefix=/global/homes/m/madams/petsc-dev/arch-xt5-opt cc="cc" doubleprecision=1": 
rm -rf build/Linux-x86_64 
rm -rf build/Linux-x86_64 
mkdir -p build/Linux-x86_64 
cd build/Linux-x86_64 && cmake /global/u2/m/madams/petsc-dev/externalpackages/metis-5.0.2 -DCMAKE_VERBOSE_MAKEFILE=1 -DGKLIB_PATH= -DCMAKE_INSTALL_PREFIX=/global/homes/m/madams/petsc-dev/arch-xt5-opt -DCMAKE_C_COMPILER=cc -DMETIS_USE_DOUBLEPRECISION=1 
cmake: /opt/gcc/4.5.3/snos/lib64/libstdc++.so.6: version `GLIBCXX_3.4.15' not found (required by cmake) 
make: *** [config] Error 1 
******************************************************************************* 

any ideas? 

Mark 




-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.mcs.anl.gov/pipermail/petsc-dev/attachments/20111130/e58f1b32/attachment.html>