Re: [AMBER-Developers] final (!?!) release candidate for Amber11

From: MengJuei Hsieh <>
Date: Mon, 19 Apr 2010 15:01:37 -0700

> From: Daniel Roe <>
> Date: Fri, 16 Apr 2010 22:29:23 -0400
> Cygwin Update:
> CYGWIN_NT-5.1 1.7.2 i686 on Athlon 64 X2 4400+; GNU 4.3.4; RC 14apr10
> ./configure -noX11 -cygwin [-mpi] gnu
> test_amber_serial: mmpbsa_py/ fails at PB step, and
> ../src/mm_pbsa/Examples/Run.mmpbsa.test fails at 02_MMPBSA_Stability, both
> with "SA Bomb in sa_arc(): Allocates aborted" messages. Perhaps these could
> be fixed like the pbsa_tsr tests (adding maxarcdot=8000 to pb namelist)?


I conclude that it's not possible to allocate that much memory in cygwin, so
I suggest we could probably change the "arcres" values in mmpbsa_py and
mm_pbsa from the default 1/16 to a coarser 1/8. But the question is, can we
still modify the programs/test cases at this point?

Although we don't encourage the use of "maxarcdot", I do have to add an
extra checking to avoid out of boundary access of those array. Here I
attached the patch.

Mengjuei Hsieh, Molecular Biology & Biochem, Univ. of California Irvine.

AMBER-Developers mailing list

Received on Mon Apr 19 2010 - 15:30:03 PDT
Custom Search