Hi,

just got a centrino laptop (for free) so have to use it whether I want it or
not . I can swear I've seen "-march=pentium-m" somewhere on the Net but
now when I actualy have the hardware and have gone through gcc man it seems
like I was hallucinating and there is now such thing.
I'd really appreciate some comments on pentium-m optimization. Has anybody
tried it with "-march=pentium4"? Any weird behavior?
AFAIK, pentium-m is just a P3 CPU with added ess2, SpeedStep gen.3, 3d
predictor, "fixed up" pipe-lines and few other tweaks. So optimizing for P3
would definitely work but seems like loosing few things, at the same time
optimizing for P4 seems a bit worrying since I'm not sure about the ABI of
both CPUs.
Comments ??

Thanks.

I believe the Pentium M arch is only in gcc 3.4. So I think you should
use Pentium 3 for march, and also use flags to enable sse2, etc.

--Geoff Mishkin <gmishkin@comcast.net>


On Wed, 09 Jun 2004 19:24:40 +0100, hrv wrote:

> Hi,
>
> just got a centrino laptop (for free) so have to use it whether I want it or
> not . I can swear I've seen "-march=pentium-m" somewhere on the Net but
> now when I actualy have the hardware and have gone through gcc man it seems
> like I was hallucinating and there is now such thing.
> I'd really appreciate some comments on pentium-m optimization. Has anybody
> tried it with "-march=pentium4"? Any weird behavior?
> AFAIK, pentium-m is just a P3 CPU with added ess2, SpeedStep gen.3, 3d
> predictor, "fixed up" pipe-lines and few other tweaks. So optimizing for P3
> would definitely work but seems like loosing few things, at the same time
> optimizing for P4 seems a bit worrying since I'm not sure about the ABI of
> both CPUs.
> Comments ??
>
> Thanks.

Geoff Mishkin wrote:

> I believe the Pentium M arch is only in gcc 3.4. So I think you should
> use Pentium 3 for march, and also use flags to enable sse2, etc.
>
> --Geoff Mishkin <gmishkin@comcast.net>

That's an option, thanks, I forgot I can add sse2 with "-msse2".
That's what I thought, gcc 3.4 should have "pentium-m" but just looked at
gcc 3.4 man and it's not to be found anywhere. Hmm!
Has anyone tried gcc 3.4? Is it usable?

hrv wrote:
> Hi,
>
> I'd really appreciate some comments on pentium-m optimization. Has anybody
> tried it with "-march=pentium4"? Any weird behavior?

I saw a page somewhere on intel.com where it is said that the gcc flag
to use for a pentium-M is 'pentium4'.

My settings :

gcc (GCC) 3.3.2 20031218 (Gentoo Linux 3.3.2-r5, propolice-3.3-7)

Linux wok 2.6.5-gentoo-r1 #7 Sun May 16 00:53:38 CEST 2004 i686 Intel(R)
Pentium(R) M processor 1700MHz GenuineIntel GNU/Linux

CFLAGS="-march=pentium4 -O2 -pipe -fomit-frame-pointer"

I had -march=pentium4 since the begining, more than 6 months now, and
never had any problem with it.

On Wed, 09 Jun 2004 21:32:29 +0200, sarylo wrote:

> hrv wrote:
>> Hi,
>>
>> I'd really appreciate some comments on pentium-m optimization. Has anybody
>> tried it with "-march=pentium4"? Any weird behavior?
>
> I saw a page somewhere on intel.com where it is said that the gcc flag
> to use for a pentium-M is 'pentium4'.
>
>
> I had -march=pentium4 since the begining, more than 6 months now, and
> never had any problem with it.

I have been installing on a P-M over the last couple days. Why I decided
to go with -march=pentium3 is the following post I found on the Kernel
list: (quote):

P-M is not a P4 core, it's an enhanced PIII core.
SSE2 was added, but compiler support for SSE2 f.p.
math shouldn't matter for the kernel.

Using P4 optimisations on a P-M may actually reduce
performance, due to the different micro-architectures.
(P4 made shifts and some leas more expensive, and
simple add/and/sub/etc less expensive.)

IOW, don't lie to the compiler and pretend P-M == P4
with that -march=pentium4.

http://www.freehackers.org/gentoo/gc...flag_gcc3.html

--
Luís Ferreira

hrv wrote:
> Hi,
>
> just got a centrino laptop (for free) so have to use it whether I want it or
> not . I can swear I've seen "-march=pentium-m" somewhere on the Net but
> now when I actualy have the hardware and have gone through gcc man it seems
> like I was hallucinating and there is now such thing.
> I'd really appreciate some comments on pentium-m optimization. Has anybody
> tried it with "-march=pentium4"? Any weird behavior?
> AFAIK, pentium-m is just a P3 CPU with added ess2, SpeedStep gen.3, 3d
> predictor, "fixed up" pipe-lines and few other tweaks. So optimizing for P3
> would definitely work but seems like loosing few things, at the same time
> optimizing for P4 seems a bit worrying since I'm not sure about the ABI of
> both CPUs.
> Comments ??
>
> Thanks.
>

Hi,

Here are the combination of flags which work the best (for me):

CFLAGS="-pipe -O3 -march=pentium4 -mmmx -msse -msse2 -mfpmath=sse,387
-maccumulate-outgoing-args -mno-align-stringops -fomit-frame-pointer
-ffast-math -funroll-all-loops -fsched-spec-load -fprefetch-loop-arrays
-ftracer -fmove-all-movables --param max-gcse-passes=4"

With gcc 3.3 It yielded the fastest code for my projects (intensive
integer calculations) and I haven't had any problems with any ebuilds at
all :-)

Regards,
numlock

Użytkownik numlock napisał:
> hrv wrote:
>
>> Hi,
>>
>> just got a centrino laptop (for free) so have to use it whether I want
>> it or
>> not . I can swear I've seen "-march=pentium-m" somewhere on the Net but
>> now when I actualy have the hardware and have gone through gcc man it
>> seems
>> like I was hallucinating and there is now such thing.
>> I'd really appreciate some comments on pentium-m optimization. Has
>> anybody
>> tried it with "-march=pentium4"? Any weird behavior?
>> AFAIK, pentium-m is just a P3 CPU with added ess2, SpeedStep gen.3, 3d
>> predictor, "fixed up" pipe-lines and few other tweaks. So optimizing
>> for P3
>> would definitely work but seems like loosing few things, at the same time
>> optimizing for P4 seems a bit worrying since I'm not sure about the
>> ABI of
>> both CPUs.
>> Comments ??
>>

Just installed 2004.2 from stage1 with
CFLAGS="-pipe -O2 -march=pentium4 -mmmx -msse -msse2
-mfpmath=sse -fomit-frame-pointer" \
running kde now and all is ok

On Thu, 26 Aug 2004 14:46:49 +0200, numlock wrote:

> hrv wrote:
>> Hi,
>>
>> just got a centrino laptop (for free) so have to use it whether I want it or
>> not . I can swear I've seen "-march=pentium-m" somewhere on the Net but
>> now when I actualy have the hardware and have gone through gcc man it seems
>> like I was hallucinating and there is now such thing.
>> I'd really appreciate some comments on pentium-m optimization. Has anybody
>> tried it with "-march=pentium4"? Any weird behavior?
>> AFAIK, pentium-m is just a P3 CPU with added ess2, SpeedStep gen.3, 3d
>> predictor, "fixed up" pipe-lines and few other tweaks. So optimizing for P3
>> would definitely work but seems like loosing few things, at the same time
>> optimizing for P4 seems a bit worrying since I'm not sure about the ABI of
>> both CPUs.
>> Comments ??
>>
>> Thanks.
>>
>
> Hi,
>
> Here are the combination of flags which work the best (for me):
>
> CFLAGS="-pipe -O3 -march=pentium4 -mmmx -msse -msse2 -mfpmath=sse,387
> -maccumulate-outgoing-args -mno-align-stringops -fomit-frame-pointer
> -ffast-math -funroll-all-loops -fsched-spec-load -fprefetch-loop-arrays
> -ftracer -fmove-all-movables --param max-gcse-passes=4"
>
> With gcc 3.3 It yielded the fastest code for my projects (intensive
> integer calculations) and I haven't had any problems with any ebuilds at
> all :-)
>
> Regards,
> numlock

Should be

"-march=pentium3"

According to the kernel developers. But as always, do whatever you want to
do .

Hi!

> "-march=pentium3"
IIRC there's a -march=pentium-m in gcc 3.4.x

--
http://tam.belchenstuermer.de/