Date: Sat, 20 Sep 1997 01:37:27 -0700 From: "Michael L. VanLoon -- HeadCandy.com" <michaelv@MindBender.serv.net> To: Mark Mayo <mark@quickweb.com> Cc: dg@root.com, Robert Schien <robsch@robkaos.ruhr.de>, hardware@freebsd.org Subject: Re: 'make world' on P6 system takes 3 h Message-ID: <199709200837.BAA22167@MindBender.serv.net> In-Reply-To: Your message of Fri, 19 Sep 97 22:32:59 -0400. <19970919223259.22704@vinyl.quickweb.com>
next in thread | previous in thread | raw e-mail | index | archive | help
>On Tue, Sep 09, 1997 at 10:06:25PM -0700, David Greenman wrote:
>> >Here are the essential options from /etc/make.conf:
>> >
>> >
>> >CFLAGS= -O2 -m486 -pipe
>>
>> Make that "-O" and kill the -m486. The -O2 nearly doubles the compile time
>> and provides almost no measurable improvement in most cases.
>
>Just curious. Did the -m486 ever really do anything?? I've always used it
>for 'make world' and kernel compiles in my 486's, but now that I'm using a
>PPro I suppose it's useless, no?
No, it's not useless. Think about it: -m486 assumes that you want
code to be optimized for processors _better_ than a 386. Why would
then removing it for a Pentium Pro be good? That means you're going
back to code that works better on a 386.
>From what I understand, most of the "benefit" from -m486 is that it
aligns data on 16-byte boundaries instead of 4 or 8-byte boundaries.
This supposedly helps cache loads happen more efficiently. Back in
"the old days", when 486s ruled, the XFree86 guys said that -m486 gave
them 10-20% improvement in speed, if I remember right. They said
speed gains were similar on Pentiums (and why shouldn't they be? --
optimizations for 486 shouldn't hurt Pentiums).
There's no reason to assume it should hurt a Pentium Pro -- in fact
going back to 4 or 8-byte alignment is almost certainly going to be
slower on the PPro, which is highly-optimized for 32-bit code, with
64-bit and 128-bit internal busses.
Supposedly there were other minor things -m486 did. It chose
instruction mixes slightly differently to prefer instructions that ran
faster on the 486, but didn't run faster on 386s. Once again, this
should, at worst, be neutral on P5s and P6s, and at best, be a slight
performance boost.
This is all from foggy memory over the last three or four years, so
take it with a grain of salt.
-----------------------------------------------------------------------------
Michael L. VanLoon michaelv@MindBender.serv.net
Contract software development for Windows NT, Windows 95 and Unix.
Windows NT and Unix server development in C++ and C.
--< Free your mind and your machine -- NetBSD free un*x >--
NetBSD working ports: 386+PC, Mac 68k, Amiga, Atari 68k, HP300, Sun3,
Sun4/4c/4m, DEC MIPS, DEC Alpha, PC532, VAX, MVME68k, arm32...
NetBSD ports in progress: PICA, others...
-----------------------------------------------------------------------------
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?199709200837.BAA22167>