Hi Roland,
I'm glad it's not just me. A link to the requested files:
https://u.pcloud.link/publink/show?code=XZF1AbXZ3Cm7SlrPOF0Wh0VR5FnMVVPB2cRk
From memory the 3D output was not enabled by default, so
QuasiLocalMeasures::output_vtk_every = $out3d_every Was not enabled. I think that variable is the only thing I changed, which is now
out3d_every = rl0_every * 2
instead of 0.
I will try to log a bug report as soon as I can. I am in the middle of a bunch of grant applications and will probably only get time this weekend.
Thanks as always for your quick and useful advice,
Chris
On Wed, 3 Feb 2021 at 04:46, Roland Haas rhaas@illinois.edu wrote:
Hello Chris,
that works perfectly.
I attach the first surface01 .vtk file, which has this weird effect.
Ok, your attached files produces bad rendering for me as well using VisIt 3.1.0.
Looking at the file (it is an ASCII file format) and comparing it to surface01_000000.vtk from the Zenodo dataset, the file is clearly different and has some very strange sections with 0.0 values in it.
The simulation that output it was from the example GW150914.
The ET gallery example parfile on linked on:
https://bitbucket.org/einsteintoolkit/einsteinexamples/raw/master/par/GW1509...
does not have vtk output enabled in QuasilocalMeasures.
Are you running the parfile from the Zenodo repo or a modified version from the gallery website? Would you mind attaching your parfile, please as well as the out and err output from Cactus?
Something strange seems to be going on in your vtk output I would say, though I have no idea what might be happening.
Having said that, I ran the gallery example parfile using the Turing (ET_2020_05, mostly b/c I had a copy already compiled) after setting:
QuasiLocalMeasures::output_vtk_every = 1
and its t=0 output for surface01_000000.vtk looks just like your bad one.
Running diff on the parameter files (from the galley page and Zenodo) the only differences I see are that WaveExtractCPM (not in the Einstein Toolkit) is disabled in the gallery version and that ADMConstraints has been replaced by ML_ADMConstraints.
You may want to try if correct output is restored after running:
cd repos/einsteinanalysis git revert 56a98ded
and recompiling.
This reverts a commit "QuasiLocalModes: Simplify VTK output routine to speed up build times" which has a suggestive commit message.
I would suggest creating a bug report ticket.
Yours, Roland
-- My email is as private as my paper mail. I therefore support encrypting and signing email messages. Get my PGP key from http://pgp.mit.edu .