From: Philip Peartree (P.Peartree_at_postgrad.manchester.ac.uk)
Date: Wed Aug 22 2007 - 10:32:50 CDT
Dow,
I've previously compiled charm++ with gcc, as I do have a mpich built with gcc
as well, but I was dumb enough to delete that. I have seen recommendations to
build with the fastest available compiler, hence I'm giving the portland
compilers a shot!
The cpu's are Opterons and the interconnect is GigE. The contents of the arch
file are below:
CMK_REAL_COMPILER=`mpiCC -show 2>/dev/null | cut -d' ' -f1 `
case "$CMK_REAL_COMPILER" in
g++) CMK_AMD64="-m64 -fPIC" ;;
esac
CMK_CPP_CHARM="/lib/cpp -P"
CMK_CPP_C="mpicc -E"
CMK_CC="mpicc $CMK_AMD64 "
CMK_CXX="mpiCC $CMK_AMD64 "
CMK_CXXPP="mpiCC -E $CMK_AMD64 "
CMK_SYSLIBS=" "
CMK_LIBS="-lckqt $CMK_SYSLIBS "
CMK_LD_LIBRARY_PATH="-Wl,-rpath,$CHARMLIBSO/"
CMK_NATIVE_CC="mpicc $CMK_AMD64 "
CMK_NATIVE_LD="mpicc $CMK_AMD64 "
CMK_NATIVE_CXX="mpiCC $CMK_AMD64 "
CMK_NATIVE_LDXX="mpiCC $CMK_AMD64 "
CMK_NATIVE_LIBS=""
# fortran compiler
CMK_CF77="mpif77"
CMK_CF90="mpif90"
CMK_F90LIBS="-L/usr/absoft/lib -L/opt/absoft/lib -lf90math -lfio -lU77
-lf77math
"
CMK_F77LIBS="-lg2c "
CMK_MOD_NAME_ALLCAPS=1
CMK_MOD_EXT="mod"
CMK_F90_USE_MODDIR=1
CMK_F90_MODINC="-p"
CMK_QT='generic64'
CMK_RANLIB="ranlib"
Thanks
Philip Peartree
Quoting Dow_Hurst <dhurst_at_mindspring.com>:
> Philip,
> I've not had trouble on mpi-linux-amd64 using gcc if the installed
> mpich was compiled with gcc.
>
> build charm++ mpi-linux-amd64 -O -nobs
>
> always gives me a working megatest or other charm++ test programs.
> What are the CPUs and interconnects you are compiling toward? That
> would be helpful at this point. You could post your charm++ arch
> file that you have modified so far.
> Dow
>
>
> -----Original Message-----
>> From: Philip Peartree <P.Peartree_at_postgrad.manchester.ac.uk>
>> Sent: Aug 22, 2007 9:35 AM
>> To: NAMD list <namd-l_at_ks.uiuc.edu>
>> Subject: namd-l: Charm++ Compilation
>>
>> Hi All,
>>
>> I'm working on a NAMD compilation, and I have compiled up the
>> Charm++ bit.. with
>> mpi-linux-amd64 (using the portland compilers) I have managed to get the
>> compilation to work using -nobs switch. The only changes I have made is to
>> remove mentions of -rdynamic from src/scripts/configure and
>> src/scripts/configure.in and to adjust the compilers used in
>> scr/arch/mpi-linux-amd64.
>>
>> I am getting a stack overflow on running the megatest program, and from what
>> I've read, it is important to get this working before embarking on
>> the actually
>> NAMD compilation.
>>
>> The error message is:
>> **ERROR: in routine alloca() there is a
>> stack overflow: thread 0, max 535822332KB, used 2KB, request -1082412208B
>>
>> Any ideas
>>
>> Philip Peartree
>> University of Manchester
>>
>
>
> No sig.
>
This archive was generated by hypermail 2.1.6 : Wed Feb 29 2012 - 15:45:08 CST