Dynamic memory allocation in scorings

Running and Runtime Errors
1

Dear FLUKA team
I am trying to understand how the dynamic memory allocation works in FLUKA-4.0.1.
I can see the hard coded USRBIN allocation has been removed, but what I find is that this can come at the expense of 1D estimators such as USRTRACK.

For example, in one run, I had USRBINS that needed ~700 MB along with 29 USRTRACK and 32 USRBDX cards as well. The input ran fine. Then I added 2 more USRTRACK cards and the the run crashed with
******* Stopped in Usrtrk: insufficient space for allocation in blank common *******
******* 22756 I*4 memory position available when 28613 needed *******

Removing one large USRBIN entry rectified the problem which suggests a hard limit some where.

My question is, what is the logic that determines this (or any other memory) limit and how is it dependent on a combination of estimators?

Thanks
Sunil

2

Hi @sunil
the automatic dynamic allocation is performed only for the memory hungry elements like USRBIN and VOXEL. All other elements are not requesting a resize of the memory if they arrive to the limit.
However with the GLOBAL card you can manually increase the memory to accommodate for everything.

Unfortunately the explanation is not on the presently distributed fluka manual. See the WHAT(6) of GLOBAL card.

Please note also that the current flair (3.1.5) is not offering this option in the GLOBAL card. It will be available in the 3.1.6 version.
So for the moment you can manually edit the card with Ctrl-E and set the WHAT(6)