AMBER 9 Bug Fixes
9.0 March, 2006 (CVS tag: v9_0)
This patch script contains all of the bug fixes detailed below in a single patch script. You can patch your amber distribution to include all of the current bug fixes by downloading this file and saving it in your $AMBERHOME directory. You should then patch your system by running:
patch -p0 -N -r patch_rejects < bugfix.all
This should patch your source files with the latest bugfixes. Any bugfixes that appear to have already been applied will be skipped. Once the patching is finished you need to recompile your copy of amber.
ptraj.c may fail to compile with some compilers.
Solvent accessibility calculations using pbsa may crash on some
Compilation of LEaP may fail on some versions of MacOSX.
Antechamber may have memory problems on some machines when analyzing ring
The frcmod.ff03 is corrupted, giving bad backbone torsions.
The "make clean" command in the lmod directories removes too many
The ambmask program has a bad initialization.
The iwrap=1 option in pmemd does not work correctly.
Poisson-Boltzmann calculations on some machines can lead to a
The top2mol2 program does not explicitly set a return value.
Minimizations in sander can trigger a divide-by-zero error.
Antechamber fails on long input files of type prepi and prepc.
bondtype may have memory problems for some molecules.
A correction (for some compilers) to bugfix 12 for Antechamber.
ptraj's actions.h may fail to compile with some compilers.
sander's lmod.f fails to compile with gfortran.
The "make clean" command in the lmod directories doesn't remove
Sander can give a bad information message in some QM/MM calculations.
Removes the requirement that there be a space after "TER" in pdb files.
Fixes quasiharmonic analysis in ptraj when the eigenvector file has
more than 9999 elements.
Workaround for an Intel v9.1 compiler bug that can cause calculations that use
divcon, including sander calculations that use divcon for the QM/MM part, to
Statistics on born radii will be incorrect if nrespai is used.
Some benchmarks that were meant only for development purposes were acidentally left in the benchmark makefile.
The ff03ua force field is missing a parameter needed for HID residues.
LEaP's setbox command ignores its third argument.
Error messages related to incorrect spin and charge combinations can give fortran formatting errors instead of the actual error message.
Increase sander's ability to handle complex masks for QM/MM
The (usused) random.F90 file in dcqtp may fail to compile with
Fix for some new C compilers that have trouble compiling Xleap.
Addresses the specific error: conflicting types for XawScrolledTableSetLocation.
Numerous minor bug fixes that may cause problems such as segfaults or
file open errors on some systems depending on the compiler being used.
Updates to the iRED analysis facility in ptraj, to provide more precision
and more eigenvalues/eigenvectors.
Fix an error when the rk2a parameter is not set for Jarzynsky runs.
Allow sander to compile when -DHAS_10_12 is specified.
Allow sander to compile on SGI Altix with netcdf support.
Fix problem with final Mulliken charges not being printed in QM/MM simulations. This patch also updates the DFTB test cases to coincide with the changes that were made by bugfix.18.
Fix problem with some gas phase simulations incorrectly switching to a less efficient list building algorithm.
Fix problem whereby some QM/MM simulations could crash if the QM region diffused out of the central box.
This bugfix resolves an issue with the SMD module in which the program would generate and error if the moving distance was large (around 100 Angstroms).
Update pmemd to provide better GB performance on the IA64 (Itanium)
The ambmask program has some obsolete use of integers as
Fix up the first energy when generalized Born is used with nrespa>1.
Add warning message and stop the calculation when spin>1 is used in QMMM calculations.
Fix to allow the "-y" switch to work for imin=5.
Fix to restart behavior in replica exchange runs.
elsize.cc may fail to compile with some compilers.
Fix possible allocation failures during sander QMMM startup.
Constant pH with closely interacting residues can misbehave due to an incorrectly formed tpair array.