Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: Deep is bad on a single processor computer?

Author: Robert Hyatt

Date: 10:02:22 10/27/03

Go up one level in this thread


On October 27, 2003 at 12:03:33, Gian-Carlo Pascutto wrote:

>On October 27, 2003 at 09:31:21, Robert Hyatt wrote:
>
>>It really isn't a "weird" reason.  These objects are persistent.  Do you want
>>all your virtual memory hogged by a process that runs amok?
>
>Shrug, this can also happen without shared mem.
>
>Windows has no problems like this.
>
>--
>GCP


Sure it does.  Try to malloc something big on a windows machine that is
installed as a server.  It has a limit for this same reason.  Otherwise
a single application can run wild and blow out virtual memory and cause
every other process to crash when they can't malloc() anything else.

I once tried to run Bruce's tablebase generator on a machine running
windows NT-server.  Wouldn't run.  Ran fine on a windows NT-client box
however.  Microsoft had a limit per-process that makes good sense.

As far as non-shared memory, you are right.  However, when a program terminates,
non-shared memory goes away.  As does memory mapped in with mmap().  Only the
system-V shared memory objects remain beyond a process's life.  Hence the
limit that is trivial to expand...




This page took 0 seconds to execute

Last modified: Thu, 15 Apr 21 08:11:13 -0700

Current Computer Chess Club Forums at Talkchess. This site by Sean Mintz.