[Caml-list] partial function application perfomance

[Caml-list] partial function application perfomance

[Arturo Borquez]

Hi all camlers,

I've being doing some benchmarks, and one of my tests
measures partial function application perfomance. I was
surprised by results: 20+ times slower than a full
function application. I expected a greater time but 20 
times is a lot more ... I've checked my program and
seems ok, indeed the results were correct. Is this
overhead normal to partial function applications?

Best regards.
Arturo


Find the best deals on the web at AltaVista Shopping!
http://www.shopping.altavista.com
-------------------
Bug reports: http://caml.inria.fr/bin/caml-bugs  FAQ: http://caml.inria.fr/FAQ/
To unsubscribe, mail caml-list-request@inria.fr  Archives: http://caml.inria.fr

Re: [Caml-list] partial function application perfomance

[Jacques Garrigue]

From: Arturo Borquez <aborquez@altavista.com>
> I've being doing some benchmarks, and one of my tests
> measures partial function application perfomance. I was
> surprised by results: 20+ times slower than a full
> function application. I expected a greater time but 20 
> times is a lot more ... I've checked my program and
> seems ok, indeed the results were correct. Is this
> overhead normal to partial function applications?

Can you post your code?
If your function is small enough, full application may result in
inlining, and a null cost. If you succeded in having your partial
application have some cost (by doing something which prohibits
inlining), the ratio may be arbitrarily large!

Jacques Garrigue
-------------------
Bug reports: http://caml.inria.fr/bin/caml-bugs  FAQ: http://caml.inria.fr/FAQ/
To unsubscribe, mail caml-list-request@inria.fr  Archives: http://caml.inria.fr