briantoby
commented
3 weeks ago It currently takes ~7 sec for GSAS-II to start on my M3 Mac (I think this goes up by a second or two when the computer has been running for a few weeks without a reboot.) There are two major areas that slow down the startup:
- Checking for the git tag # (which is the consecutively numbered release version) takes ~3 sec. This is primarily from git command (
g2repo.iter_commits('HEAD')) which could be replaced with something faster (tag,vers,gnum = g2repo.git.describe('--tags')). However, it is not clear that this is even needed, since the git tag number can be looked up from file git_version.py. The code will be changed to do that and usegit describeonly when the git_version.py file is not found. This change alone halves the startup time. - The other major time sink is in creating the menu bar contents, which requires ~2 sec of the ~3.5 needed after fixing the above. This can be reduced in two ways. One is that every data window has its own menu created for it. If we use the same menu for the data windows that have no local menu, that may provide a bit more speed at no cost. The other change is to create the menu the first time that a data tree item is used rather than to do this all before startup. This needs to be done one menu at a time, but requires relatively minor code changes.
It seems that GSAS-II is taking more time to start than it used to need. Where are the delays?