Bug/Undefined-behaviour with RESTARTing jobs

Abhishek Bagusetty abhishek... at gmail.com
Wed Nov 19 23:21:02 UTC 2014


Hi Developers,

I found some undefined-behaviour/bug in CP2K-2.5.1 with restarting jobs. 
The restart file "PROJECT.restart" has the information of the latest 
parameters overwritten at each time frame. I found that STEPS keyword which 
is the max allowed steps for a MD run is not respected. When using the 
restart file to run a job, the steps are not purged to the STEPS limit. 
Below is a code snippet from the restart file where the STEPS < 
STEP_START_VAL. 

Should a check be necessary while restarting the job to check if the 
STEP_START_VAL is not more than STEPS, so that the job is not run to 
indefinite. 

This happens for a MD run-type and may also be true for the GEO-OPT, 
CELL-OPT, etc from a quick look at the source code. Can someone please 
comment on this ? 

code snippet is given below:

********************
   &MD
     ENSEMBLE  NVE
     STEPS  20000
     TIMESTEP     9.9999999999999978E-01
     STEP_START_VAL  29640
     TIME_START_VAL     2.9640000000008175E+04
     ECONS_START_VAL    -2.1526605062223078E+02
     TEMPERATURE     3.0000000000000000E+02
     COMVEL_TOL     9.9999999999999995E-07
     ANGVEL_TOL     1.0000000000000000E-04
********************

Thanks,
Abhishek


-----------------------------------------------------------------------------------------------------------
Abhishek Bagusetty
PhD Student, Computational Modeling & Simulation
Center for Simulation and Modeling
Department of Chemical & Petroleum Engineering
University of Pittsburgh
Pittsburgh, PA - 15261
Office : 920 Benedum Hall
-----------------------------------------------------------------------------------------------------------

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.cp2k.org/archives/cp2k-user/attachments/20141119/4a23abfe/attachment.htm>


More information about the CP2K-user mailing list