From: Paul AT chocolat DOT foobar DOT co DOT uk (Paul Shirley)
Subject: Re: You can't stop progress
24 Jun 1997 12:49:56 -0700
Approved: cygnus DOT gnu-win32 AT cygnus DOT com
Distribution: cygnus
Message-ID: <$Zm6lDAPU7rzEwbJ.cygnus.gnu-win32@foobar.co.uk>
Mime-Version: 1.0
Original-Cc: gnu-win32 AT cygnus DOT com
In-Reply-To: <Pine.BSI.3.96.970623141538.21729B-100000@fish.hooked.net>
X-Mailer: Turnpike Version 3.03 <XvxOq6RVjuMF4FZcD8CYlzCnSO>
Original-Sender: owner-gnu-win32 AT cygnus DOT com

In message <Pine DOT BSI DOT 3 DOT 96 DOT 970623141538 DOT 21729B-100000 AT fish DOT hooked DOT net>,
Hmmmz <garbanzo AT hooked DOT net> writes
>I think this is mainly an OS thing and not gcc.  I've noticed that NT with
>32mb ram acts like 95 with 8, not to say that 95 is that fast, but... if
>you don't have gobs and gobs of ram at your disposal, 95 and gcc work
>together pretty well with about 32 megs of ram.
>
>- alex

Yes, this is the problem. gcc assumes that the underlying OS has good
memory management (the wasteful 2^n malloc system assumes the OS will
optimise the wastage for instance). Unfortunately Windoze seems not to
be good at managing memory. The surprise is that using >32 megs improves
gcc under windoze. I have no native win32 software that speeds up
detectably above 32M apart from gcc, I think the underlying assumptions
implicit in gnu/Unix software just stretch windoze very seriously.

-- 
Paul Shirley
-
For help on using this list (especially unsubscribing), send a message to
"gnu-win32-request AT cygnus DOT com" with one line of text: "help".