Re: threaded code in bytecode.c

[Lars Brinkhoff]

Richard Stallman <address@hidden> writes:
>     * Richard Stallman <address@hidden> at 2004-05-03 18:20-0400
>     | the main possible benefit of this change is a speedup.  How
>     | much speedup do you observe from it?  It would be good to
>     | compare various GCC optimization levels as well as the two
>     | versions of the code.
>     well, the fact is i don't know how to measure this speedup?
>     Could you suggest some agreeable method?
> You can run some computation-bound Lisp program in both versions of
> Emacs and see how long it takes.  Just call (current-time) before and
> after.  If the run time is too short to measure accurately, run the
> program 100 times.

The "Gabriel benchmarks" is a set of Lisp programs commonly used for
this purpose.  I don't think it would be too hard to port some of them
to Emacs Lisp.

ftp://ftp.cs.umass.edu/pub/eksl/gabriel/

-- 
Lars Brinkhoff,         Services for Unix, Linux, GCC, HTTP
Brinkhoff Consulting    http://www.brinkhoff.se/