[Halld-offline] Random number seeds in mcsmear

[Mark M. Ito]

David,

Don't we want to save the hdgeant random seeds so they can be used for 
another run of hdgeant later? Whey does mcsmear use the random seeds 
used by hdgeant? I would think that mcsmear needs to save its own seeds 
in case mcsmear needs to be re-run again.

Shouldn't the default remain the same as it was before: use a new random 
sequence? In other words the default should be to ignore any stored 
seeds when generating random numbers, I would think.

My 2 cents.

   -- Mark

On 01/09/2012 11:24 AM, David Lawrence wrote:
> Hi All,
>
>     I have committed changes to the repository this morning to implement
> the storage and retrieval of the random number seeds used by mcsmear.
> This required several changes:
>
> 1.) The random number generator used in mcsmear was changed from
> TRandom3 to TRandom2. This is because the former uses 624 random number
> seeds and the latter only 3. TRandom2 does have a period of 10^26 so
> this should still be sufficient.
>
> 2.) The HDDM data model was modified to include slots for the 3 seeds
> that need to be stored every event. This means old HDDM files will cause
> warnings when being read into executables linked with the new data
> model. They should still work except for the new random seed features
> may not function properly.
>
> 3.) An option "-i" was added to mcsmear to allow one to specify that any
> seeds found in the input HDDM file should be ignored. Thus, allowing one
> to re-run mcsmear using the same input file and get a different result.
> This was previously the default behavior with no option to change it.
> Note that if you are using a file with no mcsmear seeds in it (the most
> common case), specifying "-i" will have no effect and the sequence will
> be different each time you run mcsmear with that input file.
>
> Please let me know if you run into any problems with hdgeant or mcsmear
> related to these changes or if you have any questions.
>
> Regards,
> -David
>
>
>
> _______________________________________________
> Halld-offline mailing list
> Halld-offline at jlab.org
> https://mailman.jlab.org/mailman/listinfo/halld-offline


-- 
Mark M. Ito
Jefferson Lab (www.jlab.org)
(757)269-5295