[gmx-users] Gromacs 4.0.7 on Blue Gene/P

[Rongliang Wu]

Dear all,

I have installed the gromacs 4.0.7 on the Blue Gene/P according to the
instructions on the gromacs webpage, but after submitting my testing
system of water box around 10 nm, I found the simulation went really
slow, or not running at all, output files are always 0 bytes. the md.log
file gets stuck at:

----------------------------------------------------------------------
Linking all bonded interactions to atoms

The maximum number of communication pulses is: X 1 Y 1 Z 1
The minimum size for domain decomposition cells is 1.200 nm
The requested allowed shrink of DD cells (option -dds) is: 0.80
The allowed shrink of domain decomposition cells is: X 0.48 Y 0.60 Z 0.60
------------------------------------------------------------------------


I have used the double precision and fftw 3.2.2, the configure scripts
are like this for mdrun

------------------------------------------------------------------------
../configure --prefix=/home/users/wurl/software/gromacs4.0.7 --host=ppc
--build=ppc64 --enable-ppc-sqrt=1 --enable-bluegene --enable-mpi
--with-fft=fftw3 --without-x --program-suffix=-mpi-bgp   CC="mpixlc_r"
CFLAGS="-O3 -qarch=450d -qtune=450" MPICC="mpixlc_r" CXX="mpixlc_r"
CXXFLAGS="-O3 -qarch=450d -qtune=450" F77="mpixlf_r" FFLAGS="-O3
-qarch=auto -qtune=auto" LIBS="-lmass"
CPPFLAGS="-I/opt/fftw/3.2.2/include" LDFLAGS="-L/opt/fftw/3.2.2/lib"
--disable-float
------------------------------------------------------------------------

I also asked the administrator to install for me, the results were the
same.

Does anyone have similar experience?

-- 


Yours Sincerely,


Rongliang Wu
Postdoc in K.U.Leuven

--------------------------------------

Molecular and Nanomaterials Division
Department of Chemistry
Katholieke Universiteit Leuven

Celestijnenlaan 200f - bus 2404
3001 Heverlee, Belgium

+3216327456
+32488146330

rongliang.wu at chem.kuleuven.be

http://wurongliang.linkka.com/
--------------------------------------