Hi Diego--
>
> The error message that I get from running the script is:
>
> File
> "/Software/linux/64bit/xplor-nih/3.7.0.1/xplor-nih-3.7.0.1/python/simulationTools.py",
> line 1993, in runAction
> exec( s.action.replace("VALUE","%e" %s.val),
> File "<string>", line 1, in <module>
> NameError: name 'ensWeights' is not defined
your script should have:
rampedParams.append(MultRamp(2500*ensembleSize,20*ensembleSize,
"aWeights.setScale(VALUE)"))
(aWeights was written ensWweights in the script you sent earlier).
>
>
> Thanks for the advice on the ensemble size. The way I run Xplor
> (with Ne=2) is by using the command (my only machine has 20 CPUs):
>
> xplor -py -smp 16 -o saxs_struct_ensemble.out saxs_struct_ensemble.py &
> (probably not the best way).
It's probably the best way if you only have 20 cores (and you want to
save a few for other tasks). [If you specified -smp 20, the job would
complete about 17% faster.]
>
> Re chi2: I sometimes see quite large discrepancies (especially if I model a
> large number of structures). For instance below is the case for Ne=3 and
> number of structures=1000 (discrepancies occur less frequently when number or
> structure=100):
>
> saxs_struct_ensemble_3_95_0.pdb:REMARK SolnScat xray 1.800 0.002
> 3.305
> saxs_struct_ensemble_3_95_0.pdb:REMARK SolnScat xray-c 2.801 0.002
> 8.006
>
> saxs_struct_ensemble_3_228_0.pdb:REMARK SolnScat xray 1.166 0.003
> 1.388
> saxs_struct_ensemble_3_228_0.pdb:REMARK SolnScat xray-c 4.186 0.002
> 17.883
>
> Perhaps that might be the way I run Xplor?
This result means that the glob values weren't updated. Please try
adding this to the arguments of StructureLoop():
doWriteStructures=True
And you can omit the
loopInfo.writeStructure(potList)
line.
The second column of numbers above is a report of the difference of
the curves for the different ensemble members, and the small value may
indicate that the structures are more similar than you intend.
Charles
########################################################################
To unsubscribe from the XPLOR-NIH list, click the following link:
http://list.nih.gov/cgi-bin/wa.exe?SUBED1=XPLOR-NIH&A=1
