ITI increases over trials

The ITI is about 1.5sec at the start of the session, then gradually increases over trials.
At trial ~500, the ITI is around 5sec.

Is this due to the data file size getting too large? It is about 600MB for 1000 trials.

I believe this delay is in the monkeylogic ITI, not in any of the timing file code.

Do you have any suggestions where to begin looking?

What filetype are you using? if it is not BHV2, try switching to BHV2. H5 and MAT have the problem of getting slower as the filesize gets larger.

Thanks for the quick reply. I am indeed using bhv2...
I save most of the data using bhv_variable().
And some in TrialRecord.User.
Would any of those slow things down?

Of course, anything that you save with bhv_variable() slows things down. Are you saving something that grows larger each trial?

No I don't think so.
On each trial I am converting some adapters into structs and then saving them using bhv_variable(). I do this because the adapters hold relevant data.
One of them is a custom Adapter. Maybe I am not destroying all of this Adapter's graphics properly? Could this possibly lead to some accumulation of data across trials?

I have also analyzed the saved data and there is no increase in the size of data.UserVars over trials...

If it is not data saving, it can be stimulus creation that consumes the time. I guess you need to test what it is one by one. There is not so much I can do, since I do not know what your code does.

ok got it, I will run some more tests to try to isolate the problem