Re: [gcjx] Patch: FYI: parser and lexer changes

[Ranjit Mathew <rmathew at gmail dot com>]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Tom Tromey wrote:
> Ranjit> You call a 30% speedup not good enough? Whoa!
> 
> I hate to say this, since I think it is a transient condition, and I
> don't want people to really remember it, but gcjx as it is today is
> really, really slow.  It was more than 10x slower than jikes for
> building classpath; now it is merely 6x slower.

This is the classic "glass half-full v/s half-empty" case. I
think a 30% speedup in a single patch is great news, while
you consider it not being good enough to compete with Jikes.

I think you're being a bit unfair here. For an essentially
one-man project, done largely in spare time, GCJX is fairly
decent as-is. Jikes, ECJ, GCJ, etc. have had many years to
be tweaked, by many more people, sometimes working full-time.

I would personally consider correctness and robustness to
be more important at this stage than raw performance, though
I do understand that being fast would be an important criterion
for wider acceptance. "Premature performance is the root
of all evil" and all that.


> hit turns out to be pretty bad (e.g., we spend 5 seconds constructing
> format_repr objects when parsing an error-free code base).  My guess

This is weird - why do we construct these objects at all then?
Isn't this easily rectified?

Thanks,
Ranjit

- --
Ranjit Mathew      Email: rmathew AT gmail DOT com

Bangalore, INDIA.    Web: http://ranjitmathew.hostingzero.com/


-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.2 (MingW32)
Comment: Using GnuPG with Thunderbird - http://enigmail.mozdev.org

iD8DBQFDQhI0Yb1hx2wRS48RAmsqAJ94lRqS+ul1yDVq0mqDFeGygdHRmQCgllX9
kkGgf/8fBKiS4MMQwsnnRX4=
=Yx2l
-----END PGP SIGNATURE-----