run1.erl

Moderators: cyao, michael_borland

Post Reply
Max
Posts: 50
Joined: 31 Jul 2008, 12:16

run1.erl

Post by Max » 22 Oct 2010, 11:40

Dear Michael and co.,

in the Elegant Examples, ResponseMatrixFit1, in run1.erl, the file shows two columns:

&column name=ParameterValue, type=double, description="Perturbed value" &end
&column name=ParameterError, type=double, description="Perturbation value" &end

their values are the same for the FSE of the quadrupoles,
-1.282581654870910e-004 -1.282581654870910e-004 FSE PQ1 1 QUAD

but differ for the E1 and FINT items:

-4.400673214138950e-001 8.027278130105033e-003 E1 PBM 8 SBEN
3.969350296284134e-001 -2.081579208586655e-003 FINT PBM 1 SBEN

Question: what is the difference between the ParameterValue (description="Perturbed value") and the ParameterError (description="Perturbation value" ?)

Thanks, Max

michael_borland
Posts: 1959
Joined: 19 May 2008, 09:33
Location: Argonne National Laboratory
Contact:

Re: run1.erl

Post by michael_borland » 22 Oct 2010, 12:36

Max,

It is probably easiest to explain with an equation: ParameterValue = ParameterValue0 + ParameterError, where ParameterValue0 is the prior value of the parameter. If only one error is defined for a particular element and parameter, the ParameterValue0 is typically the value defined in the lattice file (or via load_parameters). If several errors are defined, you'll see them accumulate in ParameterValue.

We include both ParameterValue and ParameterError in the output file two reasons: using the ParameterError, one can check error distributions and perform correlations to get ideas about the effects of a specific error. One can also give the .erl file to load_parameters in a subsequent run to load the same perturbed values (ParameterValues) again.

--Michael

Max
Posts: 50
Joined: 31 Jul 2008, 12:16

Re: run1.erl

Post by Max » 25 Oct 2010, 15:36

Dear Michael,
thanks, it's clear now. I have another problem and I apologize for taking advantage of your patience.
I have essentially reproduced you elegantExamples of responseMatrixFit1. The sequence suggested in the example
works fine, response matrix created, etc., except that in match1.ele the program detects an error in the rpn expression:
"Problem evaluating expression: HR_M6H#1_BK#12#1.KICK -8.297321e-001 - sqr" (no other indication given).
Could you look at the problem ? I attach the match1.ele, and the .new file (in place of .lte).
Thanks a lot, Max
Attachments
run0.new
(17.76 KiB) Downloaded 835 times
match1.ele
(99.89 KiB) Downloaded 852 times

Post Reply