Re: Elisp native profiler

[Paul Eggert]

On 10/01/2012 09:34 AM, Stefan Monnier wrote:
> That's assuming that my 1ms sampling timer is really run 1000 times per
> (CPU) second.

It's true that the actual samples might come in at some slower rate,
since timers are not guaranteed to have exactly the interval
asked for.  But as long as the rate is uniform pretty much any interval
will do, as not as it's not outrageously different from what's
requested.  My guess is that it's too much trouble to add code to test for the
actual sampling rate and to try to account for the sampling-rate
error more precisely.  At any rate (no pun intended) I don't know of
any other CPU-time profiler that tries to get that fancy,
not that I'm an expert....