Message-ID: <376FCBCF.DC1921CB@uiuc.edu>
Date: Tue, 22 Jun 1999 12:45:51 -0500
From: Jon <jcmcknny AT uiuc DOT edu>
X-Mailer: Mozilla 4.6 [en] (Win98; I)
X-Accept-Language: en
MIME-Version: 1.0
To: pgcc AT delorie DOT com
Subject: Re: pgcc does better, reboot, then does terrible!(SOLVED)
References: <19990622132026 DOT QGPB8809 DOT mail DOT rdc1 DOT il DOT home DOT com AT mercury DOT snydernet DOT lan>
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Reply-To: pgcc AT delorie DOT com

Steve Snyder wrote:
> 
> On Tue, 22 Jun 1999 03:11:18 -0500, Jon wrote:
> 
> >If I use egcs for some reason -mcpu=pentiumpro -march=pentiumpro work!
> >-mstack-align-double doesn't, the compiler complains.  What works with
> >egcs1.1.2 is:
> >
> >CFLAGS = -Wall -mpentium -mcpu=pentiumpro -march=pentiumpro -O4
> >-malign-loops=2 -malign-jumps=2 -malign-functions=2 -malign-double
> >--fast-math -pg
> 
> I'm a little puzzled by your use of some of these switches.
> 
> 1. Why use -mpentium, then turn around use -mcpu=pentiumpro?  Doesn't the
> latter switch make the former redundant?

Well, I'm puzzled as to why -mcpu and -march work with egcs.  Must have
the pgcc updates in it.  Otherwise it should have errored out.

As to -mpentium.  I'm not sure how -m, -mcpu, -march are different.

For example, the kernel compiles with -m486 -DCPU=686 (or something like
that), with your CPU set to 686.  Why not use -mpentium at least??

 
> 2. Why use -O4 in preference to -O6?

As I've emailed Marc directly as a bug, with my program, using -O6 in
combo with -pg makes my program segfault.  I must go to -O4 to use -pg. 
I can  use -O6 without -pg though.

> 
> 3. Why specify the loop/jump/function alignment with the default values?
> According to the egcs doc, you should get this same alignment even if you
> dispense with the -malign-* switches entirely.

Just as a matter of convenience if I want to change them.

> 
> 4. How safe is the -malign-double switch?  The egcs doc contains a (to me)
> scary warning about breaking compatibility with "the published application
> binary interface for the 386".  Is it your experience that this is not a
> real problem?

The FAQ at pgcc's site talks about this.

Thanks,
Jon