lkml.org 
[lkml]   [1998]   [Apr]   [28]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
/
SubjectReal kernel limits
Date
From
Howdy: Is there any place where one can find out all the kernel limits that 
exists other than by munching on code for days?

Here's my situation: We just installed machines with 512M of ram, and 9G disks.
My last partition is about 5G, I'm not even sure if it can be that large.

I can easily see a time when we have 1G of Ram and 1G of swap. Something makes
me wonder if this is efficient or even possible.

If the question that comes to mind is "Do you think you need a gig of swap",
the answer is "I can certainly expect single processes to have an RSS of > 80%
of available RAM.
What is the real criteria for choosing a swap size. Back a few years ago the
suggested size was twice your RAM. This made a certain amount of sense.
But with our current siutation, I am not sure what makes sense. Our goal is
that no process should ever fail, but we hope it never swaps (because it's a
big speed hit).

In general, I am trying to figure out what the optimal configuration and setup
is, assuming that we are buying the bleeding edge of Intel boxes.

If you are wondering what we do, it's speech recognition research.
--
Hacksaw = David Charles Todd
GTE* = Hacksaw's Employer
Hacksaw's Opinions != GTE*'s Opinions
It's hard to (rec'' (e) ni'ze bee''ch)


-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu

\
 
 \ /
  Last update: 2005-03-22 13:42    [from the cache]
©2003-2011 Jasper Spaans