chicken-users
[Top][All Lists]
Advanced

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

Re: [Chicken-users] benchmarks from the Gambit world


From: Brandon J. Van Every
Subject: Re: [Chicken-users] benchmarks from the Gambit world
Date: Wed, 30 Aug 2006 09:59:59 -0700
User-agent: Thunderbird 1.5.0.5 (Windows/20060719)

felix winkelmann wrote:
On 8/30/06, Dan <address@hidden> wrote:
So, in other words, Felix tried the benchmarks pointed
to by Sven and found Chicken to be lacking... :) ?


No, not at all. Chicken's performance is just fine. I just
refuse to burn time and energy for things that are not
broken, or things that will never be quite satisfactory for
everybody, regardless how one tries.

I definitely agree about priorities. Benchmarks are synthetic priorities, useful only if they increase the marketing presence of the product. Between "it's fast" and "it works on Windows," I know what's currently more important for marketing.


For performance and numerical simulations, I found
pure Chicken unsatisfactory. Mixing in some C made it
better, but that quickly becomes limiting. Another
good option (for me) was to use OCAML with a Scheme
interpreter on top (OCS).

That's quite correct. Put metaprogramming facilities (a bit
like the stuff Will does) is IMHO the way to go.


Well, *my* philosophy is that Chicken is a BSD licensed open source project, that I can jolly well add performance optimizations to, if I ever get around to that. But I'm still working on infrastructure and I agree it comes first. It's worth noting that I picked Chicken for performance reasons once upon a time. At least, it did decently in the Shootout. It's not as fast as Bigloo, but it has a larger community and C++ support that Bigloo lacks. Bigloo has the the trick of supporting C, Java, and C#, though. I don't happen to be interested in Java and C#, though. Finally, Chicken is BSD and Bigloo is GPL. So if I get deep into compiler optimization at some point, I'm not "losing my work to the open source world." I could in principle use it as a basis for my own compiler or programming language someday, and decide what I want to do with it commercially.


Cheers,
Brandon Van Every






reply via email to

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