ufz / ogs5

OpenGeoSys THM/C simulator version 5
https://www.opengeosys.org/ogs-5
Other
39 stars 97 forks source link

Failed benchmarks #37

Closed wenqing closed 8 years ago

wenqing commented 8 years ago
norihiro-w commented 8 years ago

@wenqing There are some more CB-C benchmarks failing (https://svn.ufz.de:8443/job/OGS-5/job/ufz/branch/master/12/) after merging your PR. Can you have a look at them?

wenqing commented 8 years ago

@norihiro-w Ok.

norihiro-w commented 8 years ago

The followings can be fixed by setting the number of Gauss points to 2

norihiro-w commented 8 years ago

For fixing CB-C/Engesgaard/Kin/slow_kin_pqc/pds (PQC), the benchmark script and config type should be corrected. These two should be tested with FEM config, not PQC config, since it needs to call Phreeqc executable. PQC config uses internal Phreeqc lib.

norihiro-w commented 8 years ago

@wenqing can you also check PETSC/ConcreteCrack/decal (PETSC_GEMS)?

norihiro-w commented 8 years ago

as far as I checked, HEAD distributions are different.

wenqing commented 8 years ago

@norihiro-w As far as I could remember, Geog asked to disable PETSC/ConcreteCrack/decal. Some running errors are

 Unknown porosity model!
 Unknown diffusion model!
norihiro-w commented 8 years ago

@wenqing please check out the latest master. PETSC_GEMS config didn't work in previous versions.

wenqing commented 8 years ago

@norihiro-w It works, but the compilation needs boost being installed in advance. Just updated the result.

norihiro-w commented 8 years ago

@wenqing I don't think we can simply update the reference result for ConcreteCrack. As I wrote yesterday, there is a huge difference in hydraulic head.

wenqing commented 8 years ago

@norihiro-w Sorry I just checked the distribution of the concentration of comportments but the element velocity. Just checked the output of the element velocity of liquid/groundwater flow process, and it works good.

I have to reset the repository now.

norihiro-w commented 8 years ago

ConcreteCrack benchmark is now fixed. There was a problem in a GEMS input file.

wenqing commented 8 years ago

:+1:

norihiro-w commented 8 years ago

all T2HC and TES benchmarks are fixed.

hydromarc commented 8 years ago

The goswami-clement benchmarks are both fine in setup and reference. @norihiro-w can you please activate them again?