Hi there,
About the testing issue, we just need to add extra -Drs6000 flag in the
macosx entry of configure script, with this flag, we no longer have
strange exit value "8". Just change this line:
from
fppflags="$fppflags -DNMLEQ -DCLINK_PLAIN"
to
fppflags="$fppflags -DNMLEQ -DCLINK_PLAIN -Drs6000"
I should have put it in the fppflags. T_T
Right now it passed many test in amber8/test/ with my powerbook except
tgtmd, rdc, cnstph and whole test.sander.QMMM. (This means lmod and
divcon libraries are not doing okay for this macosx configuration.)
By the way, dacdif still thinks the difference between this two is not
acceptable. :-)
old: Ewald error estimate: 0.509E-16
new: Ewald error estimate: 0.000
Best,
--
Mengjuei Hsieh, Luo group, Molecular Biology and Biochemistry,
University of California Irvine. Tel: 49562, Address: 3144 Natural
Science I building.
Group Homepage: http://rayl0.bio.uci.edu/
On Mar 18, 2004, at 3:18 PM, MengJuei Hsieh wrote:
> By the way, just FYI, I just found that the binaries generate by this
> IBM
> compiler give exit number 8 instead of 0, so the test scripts will
> always
> goto error routine. We need to manually run ../dacdif file.save file
> instead.
Received on Wed Apr 05 2006 - 23:50:03 PDT
