> > I've run a simple test case in callgrind:
> >
> >> valgrind --tool=callgrind ./vim -U NONE -u NONE -c "echo
> >> glob('~/.vim/**/tags')" -c q
> >
> > On my current machine, I see vim_regexec() and vim_regcomp()
> > executed 14036 times, and taking respectively 41.72% + 15.67% of
> > unix_expandpath() (lr) cycles measured by callgrind.
>
> Off-topic, could you share what tool you used to get the relative
> time costs of the callgrind results? Or is that just the ratio of
> call *counts*, not wall-clock cost?
I've used kcachegrind and requested to display the percentages relative to the
parent function which has the focus. To do so I had to made sure two buttons
were pushed.
This way, in the scope of unix_expandpath() I had the ratio of cycles spent in
each callee. This may not be proportional to what would precisely happen, but
it gives a rough idea of the bottlenecks.
In another project, I took a snapshot of what kcachegrind displays:
https://wiki.orfeo-toolbox.org/index.php/File:AfterOptims-Warp.png You should
see the 2 buttons: "% Relative" and "+ Relative to parent"
--
Luc
--
--
You received this message from the "vim_dev" maillist.
Do not top-post! Type your reply below the text you are replying to.
For more information, visit http://www.vim.org/maillist.php
---
You received this message because you are subscribed to the Google Groups
"vim_dev" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to [email protected].
For more options, visit https://groups.google.com/d/optout.
