Search found 69 matches

by jan.frybort
Tue May 21, 2019 5:35 pm
Forum: Bugs, problems, questions & feeback
Topic: Bugs in version 2.1.31
Replies: 13
Views: 1102

Re: Bugs in version 2.1.31

I did the additional calculations with AMD CPU and gcc compiler for input with 100 gcu universes: *Two -omp settings, times in minutes omp 8 cmm 0 cmm 1 shbuf 0 0 18.56 21.43 shbuf 1 1 30.45 48.89 omp 16 cmm 0 cmm 1 shbuf 0 0 9.88 10.98 shbuf 1 1 23.63 40.36 *Two -omp settings, ALLOC_MEMSIZE in GB: ...
by jan.frybort
Mon May 20, 2019 8:32 pm
Forum: Bugs, problems, questions & feeback
Topic: Bugs in version 2.1.31
Replies: 13
Views: 1102

Re: Bugs in version 2.1.31

I have more findings: setting "set cmm 0" fixes both runtime and the scalability for all AMD and Intel CPUs and for both ICC and GCC, setting "set shbuf 1 1" resulted in longer runtime, runtime is bad without "set cmm 0", but scalability is OK for Intel CPUs (both ICC and GCC), runtime and scalabili...
by jan.frybort
Mon May 20, 2019 2:00 pm
Forum: Bugs, problems, questions & feeback
Topic: Bugs in version 2.1.31
Replies: 13
Views: 1102

Re: Bugs in version 2.1.31

You were right. Setting "set cmm 0" did the trick. Now the time decreased for the worst case from 19.3 min to 5.3 min. You also inspired me to experiment further and it is also dependant on OpenMP setting. Both versions (2.1.30 and 2.1.31) were compiled with the same Intel Compiler. Running the orig...
by jan.frybort
Mon May 20, 2019 12:39 pm
Forum: Bugs, problems, questions & feeback
Topic: Bugs in version 2.1.31
Replies: 13
Views: 1102

Re: Bugs in version 2.1.31

I've sent the input file via email.
Jan
by jan.frybort
Mon May 20, 2019 10:18 am
Forum: Bugs, problems, questions & feeback
Topic: Bugs in version 2.1.31
Replies: 13
Views: 1102

Re: Bugs in version 2.1.31

Dear Jaakko, thank you for releasing new Serpent version and progress of the work. I have observed problems with Serpent 2.1.31 efficiency in handling large number of universes for gcu calculations. I run 3D core calculations with hundreds of universes for calculation of macroscopic data. Here are r...
by jan.frybort
Fri Apr 27, 2018 2:33 pm
Forum: Bugs, problems, questions & feeback
Topic: Inconsistent volumes of burnable materials (solved)
Replies: 2
Views: 843

Re: Inconsistent volumes of burnable materials

Thank you, Ville. I have switched to div-card and now all the results are correct.
Jan
by jan.frybort
Thu Apr 19, 2018 1:21 am
Forum: Bugs, problems, questions & feeback
Topic: Inconsistent volumes of burnable materials (solved)
Replies: 2
Views: 843

Inconsistent volumes of burnable materials (solved)

Dear Jaakko, I am simulating fuel depletion and I am using "burn 5" card for fuel pins with Gd burnable absorber. Volume of such pins is set by vol card. I believed that the input should be the total volume of the material divided into 5 regions. This is confirmed by "-checkvolumes". But there are d...
by jan.frybort
Sun Apr 01, 2018 12:27 am
Forum: Development
Topic: Decay steps in automated burnup sequence
Replies: 15
Views: 5989

Re: Decay steps in automated burnup sequence

The depletion calculation fails with SIE: Running burnup calculation: - Step 33 / 57 - Time interval: 1 years - Burnup interval: 14.50 MWd/kgU - Cumulative burn time after step: 3.07 years - Cumulative burnup after step: 44.50 MWd/kgU - Decay step, transmutation cross sections not calculated - Batem...
by jan.frybort
Fri Mar 30, 2018 10:24 pm
Forum: Development
Topic: Decay steps in automated burnup sequence
Replies: 15
Views: 5989

Re: Decay steps in automated burnup sequence

I don't have terminal log, but according to output files it seems OK. I will try to rerun the input. It might be connected to using Stochastic Implicit Euler burnnup scheme with the following settings:

Code: Select all

set sie 10
set fsp 1 50
set cpop 10000 500 50 50
set rfw 1 
by jan.frybort
Thu Mar 29, 2018 3:09 pm
Forum: Development
Topic: Decay steps in automated burnup sequence
Replies: 15
Views: 5989

Re: Decay steps in automated burnup sequence

That list is in my message from March 27. I only omitted middle points, but otherwise the list is complete.
Jan Frybort