open TELEMAC-MASCARET The mathematically superior suite of solvers

I used the following code:
tomawac.py --merge -w oceanmesh.cas_2019-08-23-16h24min33s oceanmesh.cas

but there are some errors:
Traceback (most recent call last):
File "/home/export/online1/shij/telemac/v8p0r2/scripts/python3/tomawac.py", line 7, in <module>
main('tomawac')
File "/home/export/online1/shij/telemac/v8p0r2/scripts/python3/runcode.py", line 275, in main
run_study(cas_file, code_name, options)
File "/home/export/online1/shij/telemac/v8p0r2/scripts/python3/execution/run_cas.py", line 140, in run_study
run_local_cas(my_study, options)
File "/home/export/online1/shij/telemac/v8p0r2/scripts/python3/execution/run_cas.py", line 77, in run_local_cas
my_study.gather(options.sortie_file, options.nozip)
File "/home/export/online1/shij/telemac/v8p0r2/scripts/python3/execution/study.py", line 621, in gather
self.ncsize)
File "/home/export/online1/shij/telemac/v8p0r2/scripts/python3/execution/process.py", line 411, in process_ecr
raise TelemacException(xcpt) # raise full report
utils.exceptions.TelemacException: [{'name': 'process_ecr', 'msg': 'did not create outfile: Oceanic_201809_300s_2.r2d (WACRES)'}]

Can you try adding a flag with the number of procs you used:
--ncsize=<ncsize>

Hi Yugi,

I allowed myself to continue this thread. I use gretel.py to merge all the partitions in Cluster, while I'm wondering if it's possible to locally merge a specific number of partitions, for example, choosing 5 geographically connected pieces from all the 80 (ncsize) partitions and merge them.

Kind regards,
Yunhao

Hello Yugi,
I have a short question- Is there a way to merge the last output (NSTEPS=last) of all parallel processing results using gretel. Now gretel collects all outputs from 1 to the last NSTEPS.

I am currently using Matlab script for that. However, to get indices of mesh nodes (each core nodes indices in big mesh) takes a really huge time. Although I can save global node indices using MESH%KNOLG%I(II) in TELEMAC. But I am interested to know if gretel can already do something like that by putting somewhere NSTEPS[-1] in the code?

We thank you in advance for your support!

Ateeq

Hello,

To my knowledge, with standard gretel, the answer is no. But if you change the source code, it may be possible.

Anyway, if it helps, you can use method 2 of gretel which stores every file in memory and merge the partitionned files. It is useful when there are a lot of writings in the results files.
Please read the help: gretel.py -h (option --method 2).

Hope this helps,

Chi-Tuan