Re: Digest for scale-users-group@googlegroups.com - 1 update in 1 topic

38 views
Skip to first unread message

Larry Wetzel

unread,
Nov 21, 2025, 7:38:13 PMNov 21
to scale-us...@googlegroups.com
A few questions.

What error messages are you getting?

How many meshes are in your discrete ordinance mesh?

How much memory do you have?

What cross sections are you using?

Larry 
Sent from my iPhone

On Nov 21, 2025, at 6:30 PM, scale-us...@googlegroups.com wrote:


David Holler <dho...@lastenergy.com>: Nov 21 08:50AM -0800

Hi all,
 
I'm using MAVRIC to calculate dose responses in a reactor building using
the CAAS methodology. I have one model where I want to record the dose rate
130 feet from the source past walls and shielding, but it crashes my
computer during the first step when I run it. I typically use ~50 spatial
bins in each direction, refined at the target, and most default parameters.
I also try to reduce the total geometry volume as much as possible.
Would anyone have any suggestions or best practice points to share on these
kinds of problems in MAVRIC? Other than what is in the manual of course.
 
Thanks in advance.
You received this digest because you're subscribed to updates for this group. You can change your settings on the group membership page.
To unsubscribe from this group and stop receiving emails from it send an email to scale-users-gr...@googlegroups.com.

David Holler

unread,
Dec 16, 2025, 10:54:27 AMDec 16
to SCALE Users Group
Hi Larry,

I believe I was getting a XKBA error (apologies, this was a while ago), but it was solved by using a more powerful resource. I was initially using a laptop with 64GB memory and an AMD Ryzen 6800H CPU. The problems run without error on INL HPC's Bitterroot nodes. I see there are some parallel options for Denovo, I just haven't tried them yet.
I was using up to 50 planes in each direction, the largest being around 50 in the Y-Z plane, which has a width of about 400 meters. I'm using the 200n-47g library, with a criticality source generated in CSAS6 for the same 200n structure. I'm reading into literature on Denovo and will hopefully gain more insight on memory and CPU requirements as I go.

I had another question come up recently. I understand the Denovo scalar flux is normally an estimate in the MAVRIC sequence, used for calculating the adjoint flux and then the importance map for Monaco. Provided there are no computational limitations and the grid was refined appropriately, could the Denovo forward flux be considered an accurate representation of the scalar flux? I raise this question because I often want a full-geometry snapshot of the scalar flux without depending on Monaco. I assume the grid would have to be refined to a level that is appropriate for the Sn solver in Denovo.

Thank you for your time.

llwe...@verizon.net

unread,
Dec 16, 2025, 9:17:23 PMDec 16
to scale-us...@googlegroups.com

David,

 

I think you could use the DENOVO flux directly.  It would be like any other transport code, it would take getting the mesh and the angles to an adequate level.  I have never tried it. 

 

Larry

 

From: scale-us...@googlegroups.com <scale-us...@googlegroups.com>
Sent: Tuesday, December 16, 2025 6:31 PM
To: Digest recipients <scale-us...@googlegroups.com>
Subject: Digest for scale-us...@googlegroups.com - 1 update in 1 topic

 

Reply all
Reply to author
Forward
0 new messages