with the latest release, a new post-processing infrastructure went live. This is only the first step towards a faster, more powerful and easier to use post-processing environment.
The team is working aggressively on improving speed + capabilities and UX improvements are scheduled as well. As always, it’s important for us to know what you need most urgently to get your job done to be able to address them first. It would be great if you share with us any experiences you’ve made with the current PP environment (good and bad ones), what filters or other functionality your missing most, which workflows you’d like to see improved in which way etc.!
Well, for me the biggest problem preventing me from using the online PP is loading speed. My connection is more or less slow (5 Mbps), and maybe my geographical location has also any influence on this. I also suffer from poor graphics performance, but that could be my computer. Those two are the main reasons I download results and use Paraview.
About the filters selection, I think the “extract block” filter is useful for CFD. Also the “Surface LIC” visualization is very interesting. Being able to change to “parallel” projection would be also nice.
@dheiny
It is nice to have more option for the look of the post processing. I tried it and the setting for transparency is nice. But on the other side it is react slower than the old version. I have to be patient when I move or rotate.
The speed needs an improvement or I would go and do it offline.
@ggiraldo - loading speed is among the top priorities right now. To make sure I understand you correctly: When you talk about loading speed, do you mean the general interaction/responsiveness of the PP environment or the initial time until the first data set shows up? With the latest update (Thursday, May 19th) the general interaction should be significantly faster, specifically for larger datasets while the initial loading time was not decreased. In other words, until data shows up, it might be indeed longer than locally, but changing filters and / or timesteps should be faster than on most local machines. Then again, this was the first update with the new infrastructure and several improvements will follow shortly.
Added “Extract block” + “Surface LIC” filters to the feature request list! Parallel projection will come shortly - both for pre- and post-processing viewer.
@sburger - you’re right, as this new version uses server side rendering, there might be a bit of a lag during interaction. Working on it !
@dheiny - By loading speed I refer both to initial loading time and time to display the result of an action, like for example applying a filter, changing the time step, changing the data range or displaying the color bar.
Id like to be able to make a post processing state and be able to load it onto other simulations, this would be useful where you have many simulations with slight modifications but very similar geometry (in my case turbine blades at different wind speeds)
I’d also like to see some sort of folding method (most useful for post processing with many screenshots and states but also useful in sim design)
Id like to see a better way of orientating the view, something along the lines of traditional cad systems (yer that cube thing you can select where you look and which iso view you look at)
Now we can load multiple data sets into the post processor it would be nice to be able to translate, i.e. if I have 3 different blade design I want to see them side by side? Is that possible?
My experiences of the new post processor have been great, I love how I can now hide the bounding box. I really appreciate the opacity option and the possibilities of what I can show I are way more than I expected. In fact when I did the aerospace workshop I was like, What? when did this happen?!
As @AnnaFless said, all of your suggestions are perfectly valid and appreciated. I would like to elaborate a little on two of them:
States: they are currently linked to a specific result, I do however see the utility in being able to copy the filter pipeline and apply it to a different result. We will look into how this can be implemented for the future.
Split-view: (as in desktop ParaView) is currently not available, but I may be able to suggest a work-around for the time-being, so you can simultaneously look at two results:
– Load the solution data-sets you want to visualize, let’s say S1 and S2
– Select the data-set you would like to see translated, for example S2
– Add the transform filter to S2, and input values for the translation along the x, y or z axes
– Click on the check button to apply
– Hide The original S2 data-source
Now you have two different data-sources at two different positions. Be aware that if you choose the same field (U, p, for example) for both of them, you will be shown only one colour-bar. You may need to tweak/re-scale the value range to fit the numbers for both data-sources.
Hope this helps, and glad you hear you are enjoying your time with SimScale. Feel free to write back, should you have more questions.
Thanks, @bdaqui ill try that out! the post processing processing capability is really important in my case because I run mac and although paraview have a mac version, its incredibly unstable!
Id also like to see the ability of changing the colour scale (I can’t remember its name) where I’d be able to specify not only different colours buy opacity based on results as well, so for instance if I have an aircraft in a transient model, id like to view pressure as a cloud where the highest point of the scale is dark and gradients down to opaque (I saw this on a video somewhere of a re-entry simulation)
@bdaqui the solution for multiple results was actually what I wanted so great!
Just doing the biomedical engineering workshop it would be great if I could multi-select and alter all my filters to a different result i.e. all those transforms and reflections from von misses stress to volume stress at once?
!