Dear Steven,

I had the parallel version of mpb built for me on a large supercomputer 
and I ran some test programs. The problem that I am encountering is that 
the mpb-mpi hangs on "outputting epsilon". I looked through the 
mpb-discuss forum and I found that this is a common problem. I added the 
line

(define output-epsilon (lambda () (print "skipping output-epsilon\n")))

as you suggested to one user and in that case mpb-mpi runs fine. So it 
must be a problem with hdf5.

I saw that you suggested configuring hdf5 without mpi and mpb with 
--enable-debug. However at the same time in the mpb installation manual 
you mention that "on large supercomputers we strongly recommend using 
the parallel HDF5". So does that mean the a serial hdf5 wouldn't work at 
all on a supercomputer or just that it will be very slow? Have there 
been a workaround found for using parallel hdf5 with mpb?

What is your suggestion in this case?

Thanks,
George

_______________________________________________
mpb-discuss mailing list
[email protected]
http://ab-initio.mit.edu/cgi-bin/mailman/listinfo/mpb-discuss

Reply via email to