STATA gets much more slower after PSM

[CHANG ZHOU UZH]

I'm working on PSM, and even with same sample and code, my stata suddenly gets slower than before.
Could anyone tell me how to fix it? Should I reinstall it?

[Nick Cox]

If PSM is what I think it is, the scuttlebut is that this can be a lengthy calculation. Reinstalling Stata won't solve that;

[CHANG ZHOU UZH]

If PSM is what I think it is, the scuttlebut is that this can be a lengthy calculation. Reinstalling Stata won't solve that;

Hi Nick
Appreciate for your reply. I know propensity matching scoring takes quite time, but the propblem is it getting slower.

For example, when I work on same data and code, it took only 10 mintues before and now takes 1 hour, I'm confused with it.

[Leonardo Guizzetti]

I can think of two possible causes. One, you are downloading data from a remote source (or else the dataset is being synchronized with a backup service like Dropbox). Two, the dataset is very large and there is some element of disk caching going on which will slow things down substantially. I cannot say which, if any, are true in your case.

[CHANG ZHOU UZH]

I can think of two possible causes. One, you are downloading data from a remote source (or else the dataset is being synchronized with a backup service like Dropbox). Two, the dataset is very large and there is some element of disk caching going on which will slow things down substantially. I cannot say which, if any, are true in your case.

Hi Leonardo

Appreciate for your reply. Sry if I did't explain it clearly.

The precondidtion is that I use the same data and with same code, before and now.

It happened suddenly, it took only 10 mintues before and now takes 1 hour, I'm really confused with it.

[Leonardo Guizzetti]

The precondidtion is that I use the same data and with same code, before and now.

I understood. The code and dataset hasn't changed between runs, but the state of your computer may have. Disk caching is encountered anytime the Stata tries to hold more data in memory than the available RAM allows. If you dataset is much smaller than the typical amount of free memory, then this isn't the culprit.

Another thought occurred to me. if you are on a laptop, you may want to check that the power profile is not limiting processor speed. This can have dramatic impact on computing speed.

Beyond these ideas, I'm not in a position to further troubleshoot.

[CHANG ZHOU UZH]

I understood. The code and dataset hasn't changed between runs, but the state of your computer may have. Disk caching is encountered anytime the Stata tries to hold more data in memory than the available RAM allows. If you dataset is much smaller than the typical amount of free memory, then this isn't the culprit.

Another thought occurred to me. if you are on a laptop, you may want to check that the power profile is not limiting processor speed. This can have dramatic impact on computing speed.

Beyond these ideas, I'm not in a position to further troubleshoot.

appreciate for your advise！

I remove my stata14 and install stata16, now it seems resolve.