RatingScale really slow

lawero · April 2019

Hi!

I have this 2 projects attached.

The problem is "simple" :

the transition to the next page is very slow. I already tried with "legacy backend and set Double buffering" and "xpyriment backend and set OpenGL to 'no' " but there is no significant change. Have you any suggestion?

multi-mouse.osexp

mouse.osexp

sebastiaan · April 2019

Hi Lawero,

For me it runs fine, and I don't see anything that should make this very slow. Is OpenSesame in general very slow on your system? In that case, you might consider using a different computer for testing. Or is the slowness really specific to this experiment?

Cheers,

Sebastiaan

lawero · April 2019

mmmmm...

the experiment with 3 question doesnt require something like 2/3 seconds to you to switch page?

And the new attachment (5 questions x page) 4/5 seconds?

multi-mouse.osexp

sebastiaan · April 2019

No, other than the 1 s fixation dot in new_sketchpad, which I assume is intentional, it runs perfectly smoothly for me. Hence my recommendation to try it on another system!

lawero · April 2019

I'll try!

But for the "fixation dot in new_sketchpad" I set the duration to 1. It's 1ms or 1s?

sebastiaan · April 2019

multi-mouse-profile.osexp

Ow, right, it's indeed 1 ms! In that case it's a bit slow for me too, although not nearly as much as for you. Could you run the attached modification of your experiment? This will do a simple profiling to pinpoint where the delay comes from. It will create a profile report (profile.txt) in the same folder as the experiment file. Could you post profile.txt here?

lawero · April 2019

thx!

1-I switched system, and you are right...now the experiment timing is as yours.

2-No "profile.txt" appears. I'm using the "no-istallation version" and at the end the program reports:

"The experiment did not finish normally for the following reason: Python seems to have crashed. This should not happen. If Python crashes often, please report it on the OpenSesame forum."

(actually this happen every time but it works fine)

lawero · April 2019

help? :( any news?

sebastiaan · April 2019

Based on your previous comment, I had the impression that everything works now. (Except a Python crash as the end, but apparently with no consequences.) What is the problem that remains?

lawero · April 2019

"""Ow, right, it's indeed 1 ms! In that case it's a bit slow for me too, although not nearly as much as for you. Could you run the attached modification of your experiment? This will do a simple profiling to pinpoint where the delay comes from. It will create a profile report (profile.txt) in the same folder as the experiment file. Could you post profile.txt here?"""

And I tryed your version but no "profile.txt" appears. So I cant tell why it's slow (not so slow but still slow a little)

sebastiaan · April 2019

The profile report is probably saved somewhere else then. You could either search for it by name, or you could indicate a specific folder so that you're sure where to find the file, by changing the line that says:

with fnc.profile():

to:

with fnc.profile(r'c:\my_folder\my_profile_report.txt'):

Where of course the path should correspond to a valid path on your system.

Thanks for helping out!

lawero · April 2019

Here we go! :D Thx again

profile.txt

Any useful information? :D

sebastiaan · April 2019

Thanks! Yes, it shows that most of the time is spent rendering text. This is in itself not surprising, but it may be susceptible to performance improvements.

Thanks!

Howdy, Stranger!

Categories

RatingScale really slow

Comments

Howdy, Stranger!

Quick Links

Categories

RatingScale really slow

Comments