WHIZARD

how to control integral error goal

Asked by Si Wang

Hi,

I was running the BSM model, getting cross section integral 10^1 power of fb, while the error is 10^-3 fb.
But this error is not enough. I may want to go to 10^-5 or higher accuracy error.
How to control error to very small fb?

Thank you!
Si

Question information

Language:
English Edit question
Status:
Solved
For:
WHIZARD Edit question
Assignee:
Juergen Reuter Edit question
Solved by:
Si Wang
Solved:
Last query:
Last reply:
Revision history for this message
Juergen Reuter (j.r.reuter) said : 		#1

Hi Si,
in case you need a higher precision you have to either explicitly the number of call and adaptation steps for the Monte Carlo integration, e.g.
iterations = 6:50000:"wg", 3:100000
or you also set an error_goal for the integration:
error_goal = 5 fbarn
or
relative_error_goal = 0.001%
(Note that the numbers I put in are just proxies, but this might give you a flavor about how things work)
Cheers,
   JRR (Juergen)

Revision history for this message
Si Wang (siw34) said : 		#2

Hi Juergen,

Thanks!
That's very helpful.

Cheers,
Si