Hi there,
I'm currently writing a paper where I have to estimate several (15-20) synthetic control models, each with multiple treated units. I've used the synth_runner command previously which worked fine (only had one treated unit then), but even running a very basic version of my current model (just using a couple of covariates) is taking over an hour to run.
Just for context, I've got about 550 units, across about 50 time periods each, number of treated units varies between 20-60 depending on which model I'm estimating, I planned on using about 20 covariates when properly running my model.
Question: If a basic version of one model is taking >60mins to run, one full model I plan on running (e.g. using 20 covariates) will take significantly longer. Is there any way to speed up how Stata runs this command, or is there a command that can do the same job quicker (e.g. allsynth or others)?
Is the only solution to have them running 24/7 in the background on my uni computer and my two home laptops?
I'm currently writing a paper where I have to estimate several (15-20) synthetic control models, each with multiple treated units. I've used the synth_runner command previously which worked fine (only had one treated unit then), but even running a very basic version of my current model (just using a couple of covariates) is taking over an hour to run.
Just for context, I've got about 550 units, across about 50 time periods each, number of treated units varies between 20-60 depending on which model I'm estimating, I planned on using about 20 covariates when properly running my model.
Question: If a basic version of one model is taking >60mins to run, one full model I plan on running (e.g. using 20 covariates) will take significantly longer. Is there any way to speed up how Stata runs this command, or is there a command that can do the same job quicker (e.g. allsynth or others)?
Is the only solution to have them running 24/7 in the background on my uni computer and my two home laptops?
