synth_runner with "nested allopt" does not finish the estimation (memory problem)

[estebancolla]

Preliminaries

Before submitting an issue, please check (with x in brackets) that you:

• [x] Are using the newest release (see here for latest release version number).
• [x] Have checked that the examples in the help work.
• [x] Have read the help (HTML version).
• [x] Have checked that there is not already an existing issue for what you are reporting.
• [x] Have check running synth directly for one of the treated units works.

Expected behavior and actual behavior

I am using synth_runner with nested and allopt (from synth), and also with the option gen_vars. The routine uses a lot of working memory: 97% of 16GB of RAM memory of my computer. Sometimes it finishes the estimation; sometimes, it collapses, giving the following result:

Result when it collapses:

Estimating the treatment effects
Estimating the possible placebo effects (one set for each of the 1 treatment periods)
----+--- 1 ---+--- 2 ---+--- 3 ---+--- 4 ---+--- 5 Total: 68
...............................................x.. 2.18h elapsed. 47.12m remaining
.....xxxxxxxxxxxx| 2.70h elapsed. 
op. sys. refuses to provide memory
    In the process of initializing itself, Stata's data-storage memory manager attempted to allocate 1m bytes of
    memory and the operating system said no.
    Could not open Stata's ds3 system for merging and appending data.
r(909);

end of do-file

r(909);

Steps to reproduce the problem

(This example reproduces the use of a huge amount of memory, but does not collapse. If needed, I can provide a bigger dataset, and a routine with more predictors.)

sysuse synth_smoking, clear
tsset state year
gen byte D = (state==3 & year>=1989)
synth_runner cigsale ///
     beer(1984(1)1988) lnincome(1972(1)1988) retprice age15to24, ///
     trunit(3) trperiod(1989) ///
     nested allopt trends gen_vars
ereturn list
single_treatment_graphs
effect_graphs
pval_graphs

System information

• Stata version and flavor: Stata/SE 15.1
• OS type and version: Windows 7 Professional
• Additional hardware information
  • Memory 16 GB
  • CPU 3.20GHz
  • Hard disk: 500 GB (200 GB of free space)
• synth_runner version: updated on 5/20/2019

[bquistorff]

It looks like synth is running out of memory and failing first. synth_runner doesn't use much extra memory unless you're wanting confidence intervals in addition to p-values (so turn that off if you were using them). Not much I can do. Try a bigger machine.