Amber masthead
Filler image AmberTools19 Amber18 Manuals Tutorials Force Fields Contacts History
Filler image

Useful links:

Amber Home
Download Amber
Amber Citations
GPU Support
Intel Support
Mailing Lists
For Educators
File Formats

AmberTools Bug Fixes and Updates

Note: The individual bug fixes for AmberTools are described below in case you want to see what they do, or you don't want to use update_amber for some reason. For versions of AmberTools older than version 12, see the instructions associated with that particular version, although we strongly suggest upgrading to the latest version since numerous bugs have been fixed and they are actively supported.

Most users are encouraged to use the update_amber script (which is in $AMBERHOME, and is automatically run during the configure step) to handle patching tasks. This script automatically downloads and applies patches on this site individually and keeps track of every patch that has been applied, making patching and trouble-shooting easier.

Instructions for using update_amber can be found in the Amber Reference Manual in Chapter 2, Section 3. See Jason's wiki page for more detailed instructions. Note that update_amber has built-in Proxy support (see the Amber Reference Manual for instructions on setting up your Proxy).

If you wish to apply the listed AmberTools updates, run configure again and make sure to re-compile AmberTools (cd ${AMBERHOME} && make install).

Version 19 (April, 2019)

  • update.1: Remove some pmemd.MPI test cases that should not be run.
  • update.2: Avoid using the wrong pdb files in


  • update.3: Fix command-line parsing in


  • update.4: Add molecular reconstruction analysis to 3D-RISM calculations in

    sander, rism1d, rism3d.snglpnt

  • update.5: Have configure download the current Miniconda version (4.6.14) since the earlier version now gives problems for Python 2.
  • update.6: Fix test cases to be more resilient on newer Linux distributions.
  • update.7: Modifies


    command line argument parsing, enabling to work with 3D-RISM.
  • update.8: Modifies the way softcore free energy calculations are done in


    ; more info.
  • update.9: Add files needed to implement the ff19SB protein force field.
  • update.10: Fix the +3 and +4 ion parameters for TIP3P and TIP4P/EW water models.
  • update.11: Add a cluster size check for cluster-particle interactions in the 3D-RISM treecode.

Version 18 (April, 2018)

  • update.1: Fix compilation of


    with Intel compilers on older systems.
  • update.2: Adds an error message to


    when users run REMD and the box sizes are not the same across all replicas.
  • update.3: Fixes typo in

    's help message, and corrects E,pH,T-REMD and pH,T-REMD tests affected by Amber18's update.2.
  • update.4: Fixes test for boost libraries need by


  • update.5: Fixes a memory leak in NAB/libsff programs.
  • update.6: Updates the sample


    mdl files, to reflect up-to-date ion parameters.
  • update.7: Updates the configure script to allow cuda version 9.2 to be used.
  • update.8: This update fixes the following:
    1. Corrects REMD error trap for different box sizes in


      to account for non-periodic systems.
    2. Bugfix in


      to properly restart MultiD-REMD simulations with two or more pH, redox potential or temperature dimensions.
    3. Corrects one Amber test affected by the bugfix.
    4. Fixes typos in


  • update.9: Companion patch to the Amber update.4, updating the configure script (which is a part of AmberTools). This patch only affects


  • update.10: Fix and extend the NFE capabilities in


  • update.11: Small tweaks to the parallel version of


  • update.12: Updates test cases for mcwat in


  • update.13: Configure script changes to go along with update.12 for Amber18.
  • update.14: Bring rism1d test cases output up to the current code.

Version 17 (April, 2017)

  • update.1: Fix MDL SDF file reading and make other minor improvements in the


  • update.2: Fix the way in which


    uses gaff2.
  • update.3: Fix file format usage in


    's acdoctor mode and make other minor improvements.
  • update.4: Add option to


    to allow reduce to be used for proteins but not for ligands.
  • update.5: Fix uninitialized variable usage in


  • update.6: Add missing "cs" atom type to


  • update.7: Fix some histidine ring dihedral parameters that were incorrectly interpreted by tleap in the


    force field.
  • update.8: Add support for CUDA 9.0 compilation and add support for V100 GPUs.
  • update.9: Adds an error message to


    when users run REMD and the box sizes are not the same across all replicas.
  • update.10: Corrects REMD error trap for different box sizes in


    to account for non-periodic systems.

Version 16 (April, 2016)

  • update.1:


    , fixes typo in leaprc.water.tip4pew.
  • update.2:


    , fixes bug in RISM excess charge for solvents containing ions; bug in volume for non-orthorhomic cells; re-labels some output to be more like previous versions.
  • update.3:

    , fixes bug when trying to write a radii-adjusted prmtop file when generating a cpin file for CpHMD simulations in explicit solvent.
  • update.4: fixes apparent failure in the

    BSC1 DNA

    force field test. Code is/was always OK, but tests were using out-dated comparison files.
  • update.5: adds AmberTools components for Intel Xeon & Xeon Phi (KNL)


  • update.6: adds support for CUDA 8.0 and Pascal (GTX-1080) hardware.
  • update.7: adds a leaprc.water.spceb file, especially for use with ff15ipq
  • update.8: improve reading of ccif files in


  • update.9: New free energy routines in


    ; see nfe.pdf for full documentation.
  • update.10: Very minor tweaks to allow compilation with gcc6
  • update.11: Support Titan-XP.
  • update.12 Update


    to version 16.12. Fixes the following:
    1. Correctly handle Amber restart files with no EOL.
    2. Fix potential compilation issue with stringstreams and bleeding-edge compilers.
    3. Fix reading of force information when velocity information is not present.
    4. Fix 'stfcdiffusion' command: initial coordinates are now properly set up when atom mask selecttion does not start at the first atom, and distance calculations are now properly image when a second mask is specified.
    5. Fix 'stat' analysis command: torsion values between 0.0 and 30.0 are now correctly assigned as 'c'.
    6. Fix potential memory issue when inverting atom masks.
  • update.13: Add some missing hybridizations and non-bonded parameters to


  • update.14 Update


    to version 16.14. Fixes the following:
    1. Fix bug in 'crdaction' for actions that modify topology (e.g. strip, closest, etc).
    2. Fix reading extended format for CHARMM COR files.
    3. Fix potential segfault in 'rotdif' command when 'corrout' file cannot be opened.
    4. Fix potential segfault when reading CIF files.
    5. PDB: Atom names with 2 character element names are now correctly aligned on column 13 instead of 14.
    6. PDB: Fix conversion of nucleic acid residue names for 'pdbres'.
    7. PDB: Water residue name now changed to HOH for 'pdbres' to be consistent with PDB V3 standard.
    8. PDB: Residue numbers > 9999 and atom numbers > 99999 now wrap to 0 instead of 1 to make atom/residue counting easier.
    9. PDB: Make sure when 'terbyres' is specified that TER cards are printed when going from solvent to solute as well as vice versa.
  • update.15 Update


    to version 2.6.1. Numerous updates and improvements. For a full list, see this link for more information.
  • update.16 Update


    to version 16.16. Fixes the following:
    1. Fix compilation with OpenMPI 2.
    2. Fix handling of shape matrix information in CHARMM trajectories.
    3. Fix detection of rhombic dodecahedron unit cells.
    4. In the 'drmsd' command correctly set reference mask from reference expression.
    5. PDB: For 'terbyres' option do not put TER cards between ions (same behavior as solvent).
    6. PDB: Actually make PDB atom/residue numbers wrap to 0 after 99999/9999.
    7. PDB: For 'terbyres' option correctly output nonstandard residues as HETATM.
    8. Removes an unnecessary file from the cpptraj test directory.
  • update.17 Fix


    codes to correct for double counting of the solute-solvent potential energy.
  • update.18 Fix


    for uniform dielectric; update options when frcopt=0.
  • update.19 Fix


    to treat larger molecules, allow multiple conformations in the solvent.
  • update.20: Changes to the configure script to accommodate update.6 for Amber16.
  • update.21: Changes to the atom names and charges of the OHE residue, for use with 5' terminal phosphates in RNA.
  • update.22: Forces update_amber to use Python 2, since the upgrade process in this case requires that.
  • update.23: Adds an error message to


    when users run REMD and the box sizes are not the same across all replicas.
  • update.24: Corrects REMD error trap for different box sizes in


    to account for non-periodic systems.

Version 15 (May, 2015)

  • update.1:



    , and


    . Three main fixes:
    1. Fixes Python 2.4 and 2.5 support for and ParmEd
    2. Fixes the Python API
    3. Fixes charge scaling in cpptraj's LIE action when non-unity dielectric constant is specified
  • update.2: Adds support for CUDA 7 and fixes an erroneous error message regarding Python modules when building only OpenMP, CUDA, or CUDA-MPI versions of Amber.
  • update.3: Bugfix for LMOD docking where there are fixed atoms in the receptor
  • update.4: Update to the Lipid14 force field to add cholesterol parameters. See Madej et al, J. Phys. Chem. B. 2015.
  • update.5: Fix Python 2.6 compatibility with ParmEd
  • update.6: Adds support for CUDA 7.5. Optional support given to CUDA 7.0 due to performance regressions and GPU selection errors.

"How's that for maxed out?"

Last modified: Dec 20, 2019