Charm++ parallel programming system

["Alexandre A. Vakhrouchev" <makaveli.lcf AT gmail.com>]

I tried to build NADM on cluster with following topology:
Calculation node: 2 Intel Xeon 5365 dual core CPUs on HP Proliant BL460c
Cluster interconnection: HP Mezzanine Infiniband DDR using Voltaire
communicators
Transport network: double integrated Gigabit Ethernet

I used
>./build charm++ mpi-linux-amd --no-build-shared -O -DCMK_OPTIMIZE=1

to build CHARM++ and Linux-amd64-MPI.arch to build NAMD:

NAMD_ARCH = Linux-amd64
CHARMARCH = mpi-linux-amd64
CXX = g++
CXXOPTS = -O3 -m64 -fexpensive-optimizations -ffast-math
CC = gcc
COPTS = -O3 -m64 -fexpensive-optimizations -ffast-math

and got following warnings:

icc: Command line warning: ignoring unknown option '-m64'
icc: Command line warning: ignoring unknown option
'-fexpensive-optimizations'
icc: Command line warning: ignoring unknown option '-ffast-math'

By default MPI system is tuned to use HP MPI compilers. Thus, I guess,
changing mpiCC to mpicxx in conv-march.sh (if I spell properly) didn't work
and quit at charm++ compilation stage  with error "Unnable to use mpicxx as
linker". So I didn't change mpiCC. Than removed -lmpich from system libs
directive to successfully build NAMD (because it uses HP MPI instead of
MPICH).
NAMD works only using one cluster node. If I trie to use 2 nodes it hangs
with

Charm++ Warning: Non Charm++ Message Received.
Charm++ Warning: Non Charm++ Message Received.

Are the any keys to build CHARM++/NAMD properly for pointed cluster
configuration?
Thank you in advance!

-- 
Best regards,

Dr. Alexander Vakhrushev
Institute of Applied Mechanics
Dep. of Mech. and Phys.-Chem.
of heterogeneous mediums
UB of Russian Academy of Sciences
34 T. Baramzinoy St.
Izhevsk, Russia 426067