Question about merging MW-eABF

From: René Hafner TUK (hamburge_at_physik.uni-kl.de)
Date: Fri Feb 12 2021 - 07:08:21 CST

Dear,

    I have ran multiple setups of MW-eABF with 4 replicas.

    The namd ug states at
    https://www.ks.uiuc.edu/Research/namd/2.14/ug/node59.html#SECTION000126230000000000000
    that one has to merge the results together using the tcl script
    *eabf.tcl*

    Do understand it correctly that this is only necessary when using
    the Zheng/Yang estimator of the free energy

    but not the CZAR estimator where this is automatically done via the
    "shared on" flag or does this sharing on only reflect for the
    standard abf output files?

    It confuses me a bit.

    Checking  one of my setups with 4 replicas ran for 20.000.000 steps
    with timestep of 2fs -> in total 160ns.

     When calculating the number of counts in the replica0.zcount file I
get a sum of 19kk (~20kk, I did not use expandBoundaries here)

     While the  replica.count file results in ~80kk counts (-> already
merged results)

If I still need to merge I am missing the single czar files for the
replicas 1-3.

Are those only saved with the option `*writeCZARwindowFile **on`?*

Kind regards
René

--
Dipl.-Phys. René Hafner
TU Kaiserslautern
Germany

This archive was generated by hypermail 2.1.6 : Fri Dec 31 2021 - 23:17:10 CST