Hi,
I was unable to reproduce your error. However, I did find and fix some
uninitialized variables related to cluster averaging (purely cosmetic,
previous calculations are unaffected) that could potentially cause
strange behavior with certain compilers or on certain platforms. If
you get a chance please download the latest GitHub version
(
https://github.com/Amber-MD/cpptraj) , recompile, and see if that
fixes things.
Thanks for the report,
-Dan
On Thu, Aug 16, 2018 at 6:55 AM Charo del Genio <the.paraw.gmail.com> wrote:
>
> Dear all,
> it seems that cpptraj crashes when trying to do a clustering analysis and requesting average structures for the clusters.
> If I run the following script
>
>
> trajin ../../../Trajectories/Production.mdcrd
> cluster hieragglo epsilon 3.0 pairdist pairwisedistances.out savepairdist rms :70-74,336-340,369-374,394-402,429-434,453-458,478-482 mass summary cluster3.txt info cluster3info.txt repout cluster3rep
> avgout cluster3avg
>
>
> the output I get is
>
>
> CPPTRAJ: Trajectory Analysis. V18.01 OpenMP
> ___ ___ ___ ___
> | \/ | \/ | \/ |
> _|_/\_|_/\_|_/\_|_
>
> | Date/time: 08/16/18 11:42:46
> | Available memory: 21.034 GB
>
> Reading '../../../Topology/AFB5_dry.prmtop' as Amber Topology
> Radius Set: modified Bondi radii (mbondi)
> INPUT: Reading input from 'cluster.in'
> [trajin ../../../Trajectories/Production.mdcrd]
> Reading '../../../Trajectories/Production.mdcrd' as Amber NetCDF
> Warning: Trajectory box type is 'Trunc. Oct.' but topology box type is 'None'.
> Warning: Setting topology box information from trajectory.
> [cluster hieragglo epsilon 3.0 pairdist pairwisedistances.out savepairdist rms :70-74,336-340,369-374,394-402,429-434,453-458,478-482 mass summary cluster3.txt info cluster3info.txt repout
> cluster3rep avgout cluster3avg]
> CLUSTER: Using coords dataset _DEFAULTCRD_, clustering using RMSD (mask [:70-74,336-340,369-374,394-402,429-434,453-458,478-482]), mass-weighted best-fit
> Hierarchical Agglomerative: epsilon 3.000, average-linkage.
> Sieved frames will not be included in final cluster distance calculation.
> Pairwise distance data set is 'pairwisedistances.out'
> Saving pair-wise distances to 'pairwisedistances.out'
> Cluster information will be written to cluster3info.txt
> Summary of cluster results will be written to cluster3.txt
> Representative frames will be chosen by lowest cumulative distance to all other frames.
> Cluster representatives will be written to separate trajectories,
> prefix (cluster3rep), format Amber Trajectory
> Segmentation fault
>
>
> Note that the analysis does not even start. In fact, it does not even start reading the trajectory.
> Importantly, if I remove "avgout cluster3avg" from the script, the analysis completes without errors.
> I am using Ambertools18, compiled with gcc 7.3.0. Also, please note that I did not experience this problem with Ambertools17, suggesting the possibility that this is a bug introduced in the latest
> version.
>
>
>
> Cheers,
>
> Charo
>
>
>
>
> --
> Our new book on Synchronization phenomena is out: http://www.cambridge.org/9781107056268
>
> Dr. Charo I. del Genio
> University of Warwick
> Gibbet Hill Road
> Coventry CV4 7AL
> UK
> Tel: +44 (0)759 311 9487
>
> _______________________________________________
> AMBER mailing list
> AMBER.ambermd.org
> http://lists.ambermd.org/mailman/listinfo/amber
_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber
Received on Tue Aug 21 2018 - 08:30:03 PDT
