[Top][All Lists]

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Emacs-diffs] /srv/bzr/emacs/trunk r110342: Count overruns when prof

From: Stefan Monnier
Subject: Re: [Emacs-diffs] /srv/bzr/emacs/trunk r110342: Count overruns when profiling; change units to ns.
Date: Tue, 02 Oct 2012 22:05:34 -0400
User-agent: Gnus/5.13 (Gnus v5.13) Emacs/24.2.50 (gnu/linux)

> Which AFAIK is (was) a bug in profiler: it cannot assume that what it
> asked setitimer will actually be obeyed.  To see what interval is
> actually being loaded into the timer, one needs to call getitimer
> after setitimer returns.

If enough systems are able to return a different value in getitimer than
set in setitimer, then it would be good to do that (and signal an error
or at least output a message if the return value is different).

> I believe on Posix systems this boils down to calling clock_gettime
> with the correct timer ID (and there are equivalent APIs on
> MS-Windows).

That would be a nice approach, indeed, when CLOCK_PROCESS/THREAD_CPUTIME_ID
is supported.  We still need to worry about wraparound: counting ms
gives us a wraparound after 512M ms = 512ks =~ 5 days, which seems
acceptable, but anything finer would not fly without a good solution to
the wraparound.

> Then we could satisfy Stefan's design goal of having
> compatible profiles with different sampling period.

It's not terribly important, but it is desirable.


reply via email to

[Prev in Thread] Current Thread [Next in Thread]