Subject: Re: userid partitioned swap spaces.
To: Christos Zoulas <christos@zoulas.com>
From: John Nemeth <jnemeth@cue.bc.ca>
List: tech-kern
Date: 12/16/1998 14:15:38
On Dec 16, 10:59am, Christos Zoulas wrote:
} In article <199812160914.BAA01886@cue.bc.ca> jnemeth@cue.bc.ca (John Nemeth) writes:
} > [...] However,
} >some programs used sparse data structures, and would allocate huge
} >ammounts of memory and only use a little bit of it.  They couldn't do
} >this unless there were huge ammounts of swap space available.  Instead
} >of smacking these programs upside the head, and telling them to use
} >memory more efficiently, some "brilliant" person invented
} >overcommitting of memory (I believe this was done in SysV). 
} 
} Actually we are all talking about AIX here. The overcommitting malloc
} solution was created to address the problem of Fortran cad programs,
} that were compiled with huge arrays (because there is no dynamic memory
} allocation) and would not load to a normal workstation because it would
} run out of swap.

     Irix 4.0.5 had this problem as well.  At least it provided a
switch to turn off this "feature".  Silly question, but why are
discussing this on tech-kern since it doesn't actually pertain to
NetBSD?  What exactly is the problem with NetBSD?

}-- End of excerpt from Christos Zoulas