[Top][All Lists]
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [Chicken-users] System load
From: |
tonyg |
Subject: |
Re: [Chicken-users] System load |
Date: |
Thu, 16 Jan 2003 15:22:44 +0000 |
User-agent: |
Mutt/1.4i |
> I doubt that this is the reason for the disappointing performance.
> I'm quite sure this is more an I/O-subsystem problem...
Almost certainly. chickenlib is very basic - it barely works, let alone
works *fast* :-)
Making the core chicken scheduler work with select() is probably the
best way of making things quicker.
Sorry to all that have tried to get in touch with me re: chickenlib
etc - I've been away recently - heads down working - conceivably I'll
have a bit of time to work on things sometime in February.
> I guess this is the right moment to overhaul the scheduler...
> But that needs care. What we need is a more efficient representation
> for ready and blocked threads.
Also better integration with select() and various styles of threading
combined with I/O. (I think chickenlib polls at the moment, there's no
way of really blocking on I/O...) Careful exposure of threading
primitives would help matters a bit... perhaps
(thread-block! t)
(thread-unblock! t)
(thread-blocked? t)
or something... I can't recall what I wanted changed when I was doing
multiplex.scm, now. Damn, should have written some notes at the time.
Cheers,
Tony
--
We have to go into the jungle after them and rescue what we can. Just
remember to hang on to your sense of humor and don't let them
intimidate you.
- http://www.dourish.com/goodies/decon.html, on academics in the humanities