The Designer's Guide Community
Forum
Welcome, Guest. Please Login or Register. Please follow the Forum guidelines.
May 5th, 2024, 9:57am
Pages: 1
Send Topic Print
Running multiple montecarlo analysis from 1 file (Read 4188 times)
nano_RF
Community Member
***
Offline



Posts: 50
madison
Running multiple montecarlo analysis from 1 file
Jun 27th, 2004, 4:10am
 
Hello all,

I am trying to run multiple montecarlo simulation(with diff seed value) from same input file. I do'nt see if i can define a seed from GUI. So I manually edit the input file, and add multiple analysis. After some syntex eror i got simulation working.

My problem is i am not able to see the family curve for each MC analysis.I have given output file different name ,but still can see only one curve for each analysis.

Can anybody suggest me how can i see the family of curves?I am sure there has to be a way of seeing all the curve.My input file somewhat looks like this......

simulator lang=spectre
global 0 vdd!
include "all model file"

netlist with subcircuit definition
simulatorOptions ------

mc1 montecarlo numruns=10 seed=1 variations=process donominal=yes \
   scalarfile="../monteCarlo/mcdata1" paramfile="../monteCarlo/mcparam1" \
   saveprocessparams=yes processparamfile="../monteCarlo/processParam1" \
   processscalarfile="../monteCarlo/processData1" savefamilyplots=yes {
tran1 tran stop=350n write="spectre1.ic" writefinal="spectre1.fc".....
pss1  (  net54  net53  )  pss  fund=2.4G  harms=5  errpreset=moderate....
pnoise1  (  net54  net53  )  pnoise  sweeptype=relative....
modelParameter1 info what=models where=rawfile
element1 info what=inst where=rawfile
outputParameter1 info what=output where=rawfile
export noExprs1=oceanEval("0")
}
saveOptions1 options save=allpub

mc2 montecarlo numruns=10 seed=2 variations=process donominal=yes \
   scalarfile="../monteCarlo/mcdata2" paramfile="../monteCarlo/mcparam2" \
   saveprocessparams=yes processparamfile="../monteCarlo/processParam2" \
   processscalarfile="../monteCarlo/processData2" savefamilyplots=yes {
tran2 tran stop=350n write="spectre.ic" writefinal="spectre.fc" .....
pss2 (  net54  net53  )  pss  fund=2.4G  harms=5  errpreset=moderate...
pnoise2  (  net54  net53  )  pnoise  sweeptype=relative.....
modelParameter2 ....
}
saveOptions2 options save=allpub

--Regards,
Vikas

Back to top
 
 
View Profile   IP Logged
Andrew Beckett
Senior Fellow
******
Offline

Life, don't talk to
me about Life...

Posts: 1742
Bracknell, UK
Re: Running multiple montecarlo analysis from 1 fi
Reply #1 - Jun 27th, 2004, 10:09pm
 
Vikas,

There is a PCR filed for getting the seed value onto the monte carlo UI in ADE.

Almost certainly the problem is that the user interface doesn't  expect more than one run in the netlist - and consequently, it doesn't know what the output psf data is called, so doesn't know how to plot it.

Regards,

Andrew.
Back to top
 
 
View Profile WWW   IP Logged
nano_RF
Community Member
***
Offline



Posts: 50
madison
Re: Running multiple montecarlo analysis from 1 fi
Reply #2 - Jun 30th, 2004, 5:28pm
 
Hello Andrew,

Thanks for the reply.
But from the simlulator perspective ,I wish if one can store and later on view the results from each run.

Regards,
--Vikas
Back to top
 
 
View Profile   IP Logged
Andrew Beckett
Senior Fellow
******
Offline

Life, don't talk to
me about Life...

Posts: 1742
Bracknell, UK
Re: Running multiple montecarlo analysis from 1 fi
Reply #3 - Jun 30th, 2004, 9:40pm
 
What you could always do is to run one monte-carlo run at a time (via ADE, having hacked the netlist to change the seed), and then do a results->Save in the ADE window to save the whole run directory to another name.

Having done that you can then select a particular set of previously run results by doing results->select

BTW, why do you want to set the seed? An alternative to setting the seed might be to set the number of the first run - it means you've got the same seed, but you're further down the sequence of random numbers, which would effectively have the same effect (of course, making sure that your starting run number is greater than the number of runs in the first monte-carlo so the results don't overlap).

Regards,

Andrew.
Back to top
 
 
View Profile WWW   IP Logged
Pages: 1
Send Topic Print
Copyright 2002-2024 Designer’s Guide Consulting, Inc. Designer’s Guide® is a registered trademark of Designer’s Guide Consulting, Inc. All rights reserved. Send comments or questions to editor@designers-guide.org. Consider submitting a paper or model.