[prev in list] [next in list] [prev in thread] [next in thread] 

List:       beowulf
Subject:    Re: [Beowulf] NASTRAN on cluster
From:       Greg Lindahl <lindahl () pathscale ! com>
Date:       2005-04-12 18:53:00
Message-ID: 20050412185300.GA3958 () greglaptop
[Download RAW message or body]

On Tue, Apr 12, 2005 at 09:32:02AM -0700, Roland Krause wrote:

> Finally the whole problem is history with the arrival of 2.6.9 where
> changes to the kernel were made so that the memory spaces now grow
> towards each other from opposite sides so that one now can avoid these
> tricks allthogether, at least on 64bit machines :-) 

This isn't quite right. There are *3* things that grow: heap, mmap,
and stack. What 2.6.9 does is use your maximum stack limit to move
mmap as far from the heap as possible. So if you have an unlimited
stacksize (common for Fortran users), you still have the same old
issue. 2.6.9 is more clever for programs which use limited amounts of
stack; that's common in C programs, but not most Fortran programs.

In 64 bits, this was never an issue, because there's so much address
space available: you'll run out of physical memory long before mmap,
heap, or stack can collide.

-- greg

_______________________________________________
Beowulf mailing list, Beowulf@beowulf.org
To change your subscription (digest mode or unsubscribe) visit \
http://www.beowulf.org/mailman/listinfo/beowulf


[prev in list] [next in list] [prev in thread] [next in thread]