Search found 19 matches

by Antti Rintala
Tue Dec 03, 2019 2:52 pm
Forum: Users
Topic: Memory usage for burnup calculations
Replies: 7
Views: 235

Re: Memory usage for burnup calculations

The burnup matrixes are stored in a compressed sparse column sparse matrix format, therefore having two vectors of size nnz (one complex and one long) and one vector of size n (long), where nnz is the number of non zero entries in the burnup matrix (reactions, decays, fission yields) and n is the nu...
by Antti Rintala
Mon Nov 18, 2019 8:53 am
Forum: General
Topic: Odd error message
Replies: 1
Views: 118

Re: Odd error message

It's trying to say that there is a zero infinite spectrum few-group flux in some energy group in a universe where Serpent is trying to calculate CMM diffusion coefficients, thus causing the CMM diffusion coefficients to fail. The warning message is already made to make some more sense in the next re...
by Antti Rintala
Wed Oct 23, 2019 8:03 am
Forum: Users
Topic: Warning from function MicroCalc
Replies: 5
Views: 155

Re: Warning from function MicroCalc

If you are not using the CMM diffusion coefficients or transport cross sections (http://serpent.vtt.fi/mediawiki/index.php/Output_parameters#Diffusion_parameters) in subsequent calculations, you can disable the CMM calculation with "set cmm 0" to speed up the calculation and to prevent these error m...
by Antti Rintala
Wed Jul 03, 2019 8:16 am
Forum: Users
Topic: B1 and INF Cross-sections comparison in SFR
Replies: 6
Views: 405

Re: B1 and INF Cross-sections comparison in SFR

So are you still performing colorset calculations or are these one assembly in an infinite 2D lattice calculations? The CMM diffusion coefficients won't make any sense in the first case, and I'm not that sure that any leakage correction methods in colorset calculations makes any sense. However, ther...
by Antti Rintala
Tue Jul 02, 2019 8:25 am
Forum: Users
Topic: B1 and INF Cross-sections comparison in SFR
Replies: 6
Views: 405

Re: B1 and INF Cross-sections comparison in SFR

Did you get the zero B1_TRANSPXS using exactly which method? Did you get any error or warning messages? Are you using B1_TRANSPXS directly or are you using B1_DIFFCOEF? The old B1 method calculated the B1_TRANSPXS differently (maybe critical spectrum flux weighting of micro TRANSPXS or something lik...
by Antti Rintala
Fri Jun 28, 2019 2:53 pm
Forum: Users
Topic: Negative Homogeneous Flux when Generating ADF
Replies: 5
Views: 600

Re: Negative Homogeneous Flux when Generating ADF

Please also see this post for a possible fix to try. Note also that negative homogeneous fluxes might also be a correct solution. However, if the DF values are clearly wrong and/or have high uncertainties, then the values calculated by the built-in diffusion solver are wrong. A more correct way woul...
by Antti Rintala
Fri May 31, 2019 2:23 pm
Forum: Users
Topic: B1 and INF Cross-sections comparison in SFR
Replies: 6
Views: 405

Re: B1 and INF Cross-sections comparison in SFR

Generally it is not so simple that B1 or other fundamental spectrum group constants would be better than the infinite spectrum group constants. Are you trying to model a critical core with PARCS or the same supercell model as with Serpent? The leakage corrected group constants are maybe generally us...
by Antti Rintala
Thu May 23, 2019 7:38 am
Forum: Bugs, problems, questions & feeback
Topic: Bugs in version 2.1.31
Replies: 13
Views: 922

Re: Bugs in version 2.1.31

Thank you for performing these comparisons. The results are as expected, as the set shbuf 1 1 should basically never be used. Based on these results, as the new CMM calculation algorithm uses shared data arrays by default, we'll most probably do some changes to default CMM calculation options.
by Antti Rintala
Tue May 21, 2019 8:27 am
Forum: Bugs, problems, questions & feeback
Topic: Bugs in version 2.1.31
Replies: 13
Views: 922

Re: Bugs in version 2.1.31

Could you still try the same inputs using "set shbuf 0 0" instead of "set shbuf 1 1" with both "set cmm 0" and without it? This should lead into larger memory consumption, so depending on your memory resources this option may not be possible to be used.
by Antti Rintala
Thu Jan 19, 2017 11:16 am
Forum: Input
Topic: Albedos with set alb
Replies: 1
Views: 4571

Re: Albedos with set alb

Hi, I've been recently calculating some albedos but the hexagonal core lattice was of the y type whereas yours is of the x type. I assume ALB_IN_CURR are in-currents in 2-groups for 6 sides of hexagon, but in which order? I have checked this for the y type hexagon, and it seems that the ordering beg...