======================================================================
INSTRUCTIONS FOR INSTALLING PYLITH ON THE CALTECH LINUX CLUSTER PANGU
======================================================================

These instructions are nearly identical to those in
Cluster_CIT-garuda.txt. As a result, we only mention those settings
which differ.
 
See the files in the Cluster_CIT-pangu directory for the settings used
to setup the environment variables using the bash shell. The
$TOOLS_FORMAT environment variable is used to differentiate between
packages built with different compilers and/or different optimization
settings. In this case, TOOLS_FORMAT is intel-11.1. Because this is a
cluster the tools are built for production runs using
high-optimization level. You should setup your environment BEFORE
building the packages.

We use the Intel 11.1 compiler and MKL math libraries. These are in
  /opt/intel/Compiler/11.1/072/
  /opt/intel/Compiler/11.1/072/mkl/lib

We use OpenMPI version 1.4.2 installed in
  /opt/openmpi/1.4.2/intel


netcdf

  Run configure w/netcdf-4 disabled.

    $HOME/src/netcdf-4.1.1/netcdf-4.1.1/configure --prefix=${TOOLS_DIR} --enable-shared --disable-netcdf-4 CC=icc CXX=icpc FC=ifort F77=ifort CFLAGS="-O2 -fPIC" CXXFLAGS="-O2 -fPIC" FFLAGS="-O2 -fPIC" FCFLAGS="-O2 -fPIC"


numpy

  Edit site.cfg to use the Intel KML library version 11.1.
    [mkl]
    library_dirs = /opt/intel/Compiler/11.1/072/mkl/lib/em64t
    lapack_libs = mkl_lapack
    mkl_libs = mkl_core, mkl_intel_thread, guide

PETSc

  Run configure and use Intel MKL library version 11.1.

    config/configure.py --PETSC_ARCH=${TOOLS_FORMAT}_opt --with-debugging=0 --with-clanguage=c++ --with-mpi-compilers=1 --with-shared=1 --with-dynamic=1 --with-64-bit-points=1 --with-large-file-io=1 --with-blas-lapack-dir=/opt/intel/Compiler/11.1/072/mkl/lib/em64t --with-lgrind=0 -with-chaco=1 --download-chaco=1 --with-parmetis=1 --download-parmetis=1 --with-sieve=1 --with-boost=1 --download-boost=1 --with-ml=1 --download-ml=1
