First time here? Check our help page!
3

Easiest way to run multiple OpenStudio simulations

What is the easiest way to run multiple OpenStudio simulations on your own desktop. I have used the Large Scale Analysis Tool for heavy lifting of thousands of simulations, but if I just want to run 10 simulations on my own machine, what is the best way to do this? Essentially, I want the same functionality as the Run Multiple IDFs capability in EP Launch were you can run a simulation on each core.

dpud12's avatar
1.3k
dpud12
asked 2017-03-15 19:35:32 -0500
__AmirRoth__'s avatar
4.4k
__AmirRoth__
updated 2017-04-01 15:36:51 -0500
edit flag offensive 0 remove flag close merge delete

Comments

add a comment see more comments

1 Answer

2

OpenStudio's Parametric Analysis Tool. Design alternatives can be constructed using measures or they can be individual models.

ljbrackney's avatar
3.8k
ljbrackney
answered 2017-03-16 09:58:21 -0500, updated 2017-03-16 10:01:41 -0500
edit flag offensive 0 remove flag delete link

Comments

@ljbrackney The PAT works great if you have separate models without any measures attached, but in my case I have 15 different models, each with unique measures. When I input the models into PAT, it strips the OSMs of the measures that I have assigned. I see that there is an option to add a User Defined measure after inputting an alternative starting model, but when I do this it doesn't allow me to select a measure from the ones I have created and enter the relevant inputs . . . I must be missing something cause I would think PAT would want to support this functionality

dpud12's avatar dpud12 (2017-03-16 10:53:58 -0500) edit

Ah. Yeah, PAT 1.x doesn't support that. I believe PAT 2.1.0 does, although I haven't been testing multiple seed model functionality personally. @aparker or @David Goldwasser are probably in a better position to speak to that capability - at least until I get around to testing and documenting it.

ljbrackney's avatar ljbrackney (2017-03-16 11:14:39 -0500) edit

Yes PAT 2.x does support custom seed models for each datapoint. This would allow you to build up an analysis of unique seed models with with the same or custom measures run on them. Another option I thought of when I saw this post was using the OpenStudio Command Line Interface, but for most use cases I think PAT 2.1 will be the way to go. 2.1 is due out at the end of March. The 1.14.0 release of PAT is the last 1.x PAT that will be released.

David Goldwasser's avatar David Goldwasser (2017-03-16 12:11:57 -0500) edit

@ljbrackney and @David Goldwasser thanks for the insight, looking forward to PAT 2.1 and glad that it will solve this issue

dpud12's avatar dpud12 (2017-03-20 13:05:41 -0500) edit
add a comment see more comments