Computer Chess Club Archives




Subject: Re: POSIX threads

Author: Steven Edwards

Date: 00:52:40 07/06/05

Go up one level in this thread

On July 05, 2005 at 21:39:48, Lance Perkins wrote:

>You missed the point here. I never claimed that a shared resource is a single
>memroy. Read again. Even on a single CPU you would need locks, mutexes,
>semaphores, events, mailboxes, etc to synchronize. But to say that all you need
>is a single-cpu to "test" your multi-threaded app is too naive of a statement.

If every shared resource is properly guarded by a mutex, then there will not be
thread faults due to race conditions.  This is true as true on a single CPU box
as it is on a 64 CPU box.

The only time a dual CPU box will encounter race where a single CPU box won't is
when both CPUs hit the same memory address at the same time.  But this can't
happen if the memory address is part of a shared resource properly guarded by a
mutex.  All other possible thread faults can be tested on a single CPU machine,
although they might be fould faster on a multi CPU machine.

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.