This is the mail archive of the gcc@gcc.gnu.org mailing list for the GCC project.

Index Nav:	[Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav:	[Date Prev] [Date Next]	[Thread Prev] [Thread Next]
Other format:	[Raw text]

Re: Tremendous increase in compile times for 3.4.1 with -mcpu=G5

From: "Eric Botcazou" <ebotcazou at libertysurf dot fr>
To: "Dale Johannesen" <dalej at apple dot com>
Cc: "Bradley Lucier" <lucier at math dot purdue dot edu>,"Matt Austern" <austern at apple dot com>,<gcc at gcc dot gnu dot org>,"Michael Matz" <matz at suse dot de>,"David Edelsohn" <dje at watson dot ibm dot com>
Date: Sun, 15 Aug 2004 15:35:09 +0200
Subject: Re: Tremendous increase in compile times for 3.4.1 with -mcpu=G5
References: <C7AB5C7A-EE24-11D8-8F7A-003065BA681E@math.purdue.edu> <200408142032.i7EKWCD29304@makai.watson.ibm.com> <D3AA48B7-EE32-11D8-9C4D-000A95D7D16C@apple.com>

> I believe Sparc V8plus mode is the only thing similar.  That might well
> have the same problem.

I would not think so.  SPARC V8+ is a much more limited mixed mode than PPC
G5 as far as I know.  And my understanding is that, in this mode, the
back-end does 64-bit computation behind the back of the middle-end, in
particular of the register allocator which still thinks everything is
32-bit.  And of course, as rth would put it, lie to reload and it will have
its revenge (e.g. PR target/12958).

--
Eric Botcazou

Follow-Ups:
- Re: Tremendous increase in compile times for 3.4.1 with -mcpu=G5
  - From: Dale Johannesen

References:
- Re: Tremendous increase in compile times for 3.4.1 with -mcpu=G5
  - From: Bradley Lucier
- Re: Tremendous increase in compile times for 3.4.1 with -mcpu=G5
  - From: David Edelsohn
- Re: Tremendous increase in compile times for 3.4.1 with -mcpu=G5
  - From: Dale Johannesen

Index Nav:	[Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav:	[Date Prev] [Date Next]	[Thread Prev] [Thread Next]