public inbox for gentoo-dev@lists.gentoo.org
 help / color / mirror / Atom feed
* [gentoo-dev] Warning about -march=k6*
@ 2003-07-14  6:14 Mike Frysinger
  2003-07-14 11:21 ` Tim Ruehsen
                   ` (4 more replies)
  0 siblings, 5 replies; 7+ messages in thread
From: Mike Frysinger @ 2003-07-14  6:14 UTC (permalink / raw
  To: gentoo-dev

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

i have seen many bugs where people who have used -march=k6* in their CFLAGS 
experienced segfaults, compile failures, etc...

i'd like to propose adding a warning to make.conf to deter ppl from using 
these flags ... rather i'd suggest they use -mcpu=k6* ...

the bug can be found at http://bugs.gentoo.org/show_bug.cgi?id=24379
i just want to know if anyone out there has had *success* with this flag ...
- -mike
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.2 (GNU/Linux)
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=Tvtp
-----END PGP SIGNATURE-----

--
gentoo-dev@gentoo.org mailing list


^ permalink raw reply	[flat|nested] 7+ messages in thread

* Re: [gentoo-dev] Warning about -march=k6*
  2003-07-14  6:14 [gentoo-dev] Warning about -march=k6* Mike Frysinger
@ 2003-07-14 11:21 ` Tim Ruehsen
  2003-07-14 14:27 ` Brad Cowan
                   ` (3 subsequent siblings)
  4 siblings, 0 replies; 7+ messages in thread
From: Tim Ruehsen @ 2003-07-14 11:21 UTC (permalink / raw
  To: vapier, gentoo-dev

Yes, I had that problem which ceased after changing architecture from i686 to 
i586.

Tim

Am Montag, 14. Juli 2003 08:14 schrieb Mike Frysinger:
> i have seen many bugs where people who have used -march=k6* in their CFLAGS
> experienced segfaults, compile failures, etc...
>
> i'd like to propose adding a warning to make.conf to deter ppl from using
> these flags ... rather i'd suggest they use -mcpu=k6* ...
>
> the bug can be found at http://bugs.gentoo.org/show_bug.cgi?id=24379
> i just want to know if anyone out there has had *success* with this flag
> ... -mike


--
gentoo-dev@gentoo.org mailing list


^ permalink raw reply	[flat|nested] 7+ messages in thread

* Re: [gentoo-dev] Warning about -march=k6*
  2003-07-14  6:14 [gentoo-dev] Warning about -march=k6* Mike Frysinger
  2003-07-14 11:21 ` Tim Ruehsen
@ 2003-07-14 14:27 ` Brad Cowan
  2003-07-14 18:32 ` C. Brewer
                   ` (2 subsequent siblings)
  4 siblings, 0 replies; 7+ messages in thread
From: Brad Cowan @ 2003-07-14 14:27 UTC (permalink / raw
  To: gentoo-dev

On Mon, 14 Jul 2003 02:14:32 -0400
Mike Frysinger <vapier@gentoo.org> wrote:

> -----BEGIN PGP SIGNED MESSAGE-----
> Hash: SHA1
> 
> i have seen many bugs where people who have used -march=k6* in their
> CFLAGS experienced segfaults, compile failures, etc...
> 
> i'd like to propose adding a warning to make.conf to deter ppl from
> using these flags ... rather i'd suggest they use -mcpu=k6* ...
> 

Most all of the problems I've seen are because the user had a CHOST of
i686 instead of i586, or used options like -funroll-loops which break
things on this arch. Also some use the wrong k6 option ie k6-3 when they
have a k6-2 etc.., but I definately would advize at least a big fat
warning. I would say go with -march=i586 instead of -mcpu=k6 if you want
to go that route.

-- 
Brad Cowan <bcowan@gentoo.org>
Developer,
Gentoo Linux	http://www.gentoo.org/~bcowan

Public Key: http://pgp.mit.edu:11371/pks/lookup?op=get&search=0xB1F16A56
Key fingerprint = C408 75B9 E68D 26E2 EAAE  20CF 4D5E 293D B1F1 6A56

--
gentoo-dev@gentoo.org mailing list


^ permalink raw reply	[flat|nested] 7+ messages in thread

* Re: [gentoo-dev] Warning about -march=k6*
  2003-07-14  6:14 [gentoo-dev] Warning about -march=k6* Mike Frysinger
  2003-07-14 11:21 ` Tim Ruehsen
  2003-07-14 14:27 ` Brad Cowan
@ 2003-07-14 18:32 ` C. Brewer
  2003-07-14 18:32 ` Stewart Honsberger
  2003-07-15  1:07 ` Brad Laue
  4 siblings, 0 replies; 7+ messages in thread
From: C. Brewer @ 2003-07-14 18:32 UTC (permalink / raw
  To: gentoo-dev

[-- Attachment #1: Type: text/plain, Size: 1146 bytes --]

On Mon, 14 Jul 2003 02:14:32 -0400
Mike Frysinger <vapier@gentoo.org> wrote:

> the bug can be found at http://bugs.gentoo.org/show_bug.cgi?id=24379
> i just want to know if anyone out there has had *success* with this flag
> ...

I've had much success with this flag, after long trying times with learning
what flags you can get away with in regards to the different packages. I
first started with-
-O3 -march=k6-2 -pipe -funroll-loops -fomit-frame-pointer (continuing list
of -f functions)
and whittled down to-
-O2 -march=k6-2 -pipe -fomit-frame-pointer with these observations-

-O3 and -funroll-loops work together if not appended with -fPIC or any other
-*PIC option by the source.
-O3 and -funroll-loops will build compression packages (bzip2,jpeg,etc) but
will not allow them to execute correctly.
Bottom line is that since gcc 3.x, the difference between -O3 and -O2 on
k6's speedwise is imperceptible to the end user, and I advise really staying
away from -O3 and excessive -f* functions, especially in the loop dept.:)
-- 
Chuck Brewer
Registered Linux User #284015
Get my gpg public key at pgp.mit.edu!! Encrypted e-mail preferred.



[-- Attachment #2: Type: application/pgp-signature, Size: 189 bytes --]

^ permalink raw reply	[flat|nested] 7+ messages in thread

* Re: [gentoo-dev] Warning about -march=k6*
  2003-07-14  6:14 [gentoo-dev] Warning about -march=k6* Mike Frysinger
                   ` (2 preceding siblings ...)
  2003-07-14 18:32 ` C. Brewer
@ 2003-07-14 18:32 ` Stewart Honsberger
  2003-07-14 18:50   ` Jon Portnoy
  2003-07-15  1:07 ` Brad Laue
  4 siblings, 1 reply; 7+ messages in thread
From: Stewart Honsberger @ 2003-07-14 18:32 UTC (permalink / raw
  To: vapier; +Cc: gentoo-dev

Mike Frysinger wrote:
> i have seen many bugs where people who have used -march=k6* in their CFLAGS
> experienced segfaults, compile failures, etc...
> 
> i'd like to propose adding a warning to make.conf to deter ppl from using
> these flags ... rather i'd suggest they use -mcpu=k6* ...
> 
> the bug can be found at http://bugs.gentoo.org/show_bug.cgi?id=24379
> i just want to know if anyone out there has had *success* with this flag ...

My entire server (AMD K62-400) is built around -march=k6; stage1 
ground-up, however I am using the 1.2 profile with GCC 2.95.

Does -march=k6-2 exist for 2.95? I'd hate to destabalize my server to 
test, but I might be able to scrounge up another of the same chip and 
motherboard.

-- 
Stewart Honsberger
Gentoo Developer
http://www.snerk.org/


--
gentoo-dev@gentoo.org mailing list


^ permalink raw reply	[flat|nested] 7+ messages in thread

* Re: [gentoo-dev] Warning about -march=k6*
  2003-07-14 18:32 ` Stewart Honsberger
@ 2003-07-14 18:50   ` Jon Portnoy
  0 siblings, 0 replies; 7+ messages in thread
From: Jon Portnoy @ 2003-07-14 18:50 UTC (permalink / raw
  To: Stewart Honsberger; +Cc: vapier, gentoo-dev

On Mon, Jul 14, 2003 at 02:32:45PM -0400, Stewart Honsberger wrote:
> Mike Frysinger wrote:
> >i have seen many bugs where people who have used -march=k6* in their CFLAGS
> >experienced segfaults, compile failures, etc...
> >
> >i'd like to propose adding a warning to make.conf to deter ppl from using
> >these flags ... rather i'd suggest they use -mcpu=k6* ...
> >
> >the bug can be found at http://bugs.gentoo.org/show_bug.cgi?id=24379
> >i just want to know if anyone out there has had *success* with this flag 
> >...
> 
> My entire server (AMD K62-400) is built around -march=k6; stage1 
> ground-up, however I am using the 1.2 profile with GCC 2.95.
> 
> Does -march=k6-2 exist for 2.95? I'd hate to destabalize my server to 
> test, but I might be able to scrounge up another of the same chip and 
> motherboard.

Not AFAIK.

> 
> -- 
> Stewart Honsberger
> Gentoo Developer
> http://www.snerk.org/
> 
> 
> --
> gentoo-dev@gentoo.org mailing list

-- 
Jon Portnoy
avenj/irc.freenode.net

--
gentoo-dev@gentoo.org mailing list


^ permalink raw reply	[flat|nested] 7+ messages in thread

* Re: [gentoo-dev] Warning about -march=k6*
  2003-07-14  6:14 [gentoo-dev] Warning about -march=k6* Mike Frysinger
                   ` (3 preceding siblings ...)
  2003-07-14 18:32 ` Stewart Honsberger
@ 2003-07-15  1:07 ` Brad Laue
  4 siblings, 0 replies; 7+ messages in thread
From: Brad Laue @ 2003-07-15  1:07 UTC (permalink / raw
  To: gentoo-dev

Mike Frysinger wrote:

> the bug can be found at http://bugs.gentoo.org/show_bug.cgi?id=24379
> i just want to know if anyone out there has had *success* with this flag ...

I'm using a K6-2 3D with the following attributes:

processor       : 0
vendor_id       : AuthenticAMD
cpu family      : 5
model           : 8
model name      : AMD-K6(tm) 3D processor
stepping        : 12

I'm using -march=k6-2 without problems.

Brad


--
gentoo-dev@gentoo.org mailing list


^ permalink raw reply	[flat|nested] 7+ messages in thread

end of thread, other threads:[~2003-07-15  1:07 UTC | newest]

Thread overview: 7+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2003-07-14  6:14 [gentoo-dev] Warning about -march=k6* Mike Frysinger
2003-07-14 11:21 ` Tim Ruehsen
2003-07-14 14:27 ` Brad Cowan
2003-07-14 18:32 ` C. Brewer
2003-07-14 18:32 ` Stewart Honsberger
2003-07-14 18:50   ` Jon Portnoy
2003-07-15  1:07 ` Brad Laue

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox