* Optimization for PII
@ 1999-09-13 2:35 Rainer Dorsch
1999-09-13 13:00 ` Marc Lehmann
1999-09-30 18:02 ` Rainer Dorsch
0 siblings, 2 replies; 10+ messages in thread
From: Rainer Dorsch @ 1999-09-13 2:35 UTC (permalink / raw)
To: gcc
The egcs documentation says
`-march=CPU TYPE'
Generate instructions for the machine type CPU TYPE. The choices
for CPU TYPE are: `i386', `i486', `pentium', and `pentiumpro'.
Specifying `-march=CPU TYPE' implies `-mcpu=CPU TYPE'.
It does not mention PII and PIII. Does pentiumpro run on PII and PIII or is
there another undocumented option (I am using gcc-2.95.1)?
Thanks.
--
Rainer Dorsch
Abt. Rechnerarchitektur e-mail:rainer.dorsch@informatik.uni-stuttgart.de
Uni Stuttgart Tel.: 0711-7816-215
^ permalink raw reply [flat|nested] 10+ messages in thread
* Re: Optimization for PII
1999-09-13 2:35 Optimization for PII Rainer Dorsch
@ 1999-09-13 13:00 ` Marc Lehmann
1999-09-30 18:02 ` Marc Lehmann
1999-09-30 18:02 ` Rainer Dorsch
1 sibling, 1 reply; 10+ messages in thread
From: Marc Lehmann @ 1999-09-13 13:00 UTC (permalink / raw)
To: gcc
On Mon, Sep 13, 1999 at 11:34:45AM +0300, Rainer Dorsch <rainer@rainer.informatik.uni-stuttgart.de> wrote:
>
> The egcs documentation says
>
> `-march=CPU TYPE'
> Generate instructions for the machine type CPU TYPE. The choices
> for CPU TYPE are: `i386', `i486', `pentium', and `pentiumpro'.
> Specifying `-march=CPU TYPE' implies `-mcpu=CPU TYPE'.
>
> It does not mention PII and PIII. Does pentiumpro run on PII and PIII or is
> there another undocumented option (I am using gcc-2.95.1)?
i386 == i486 == pentium == k6 < pentiumpro == pii == piii
(e.g. i386 code runs on pii, but pii code not on i386)
This is the view the compiler currently has. The pii and piii additional
instructions are not yet used, as isn't mmx.
It is unclear (to me) wether this will change, as there is no real strong
hierarchy anymore (chip can have mmx but not fcomi, or vice versa), or
wether additional switches (-mmx, -m3dnow) will be introduced.
--
-----==- |
----==-- _ |
---==---(_)__ __ ____ __ Marc Lehmann +--
--==---/ / _ \/ // /\ \/ / pcg@goof.com |e|
-=====/_/_//_/\_,_/ /_/\_\ XX11-RIPE --+
The choice of a GNU generation |
|
^ permalink raw reply [flat|nested] 10+ messages in thread
* Optimization for PII
1999-09-13 2:35 Optimization for PII Rainer Dorsch
1999-09-13 13:00 ` Marc Lehmann
@ 1999-09-30 18:02 ` Rainer Dorsch
1 sibling, 0 replies; 10+ messages in thread
From: Rainer Dorsch @ 1999-09-30 18:02 UTC (permalink / raw)
To: gcc
The egcs documentation says
`-march=CPU TYPE'
Generate instructions for the machine type CPU TYPE. The choices
for CPU TYPE are: `i386', `i486', `pentium', and `pentiumpro'.
Specifying `-march=CPU TYPE' implies `-mcpu=CPU TYPE'.
It does not mention PII and PIII. Does pentiumpro run on PII and PIII or is
there another undocumented option (I am using gcc-2.95.1)?
Thanks.
--
Rainer Dorsch
Abt. Rechnerarchitektur e-mail:rainer.dorsch@informatik.uni-stuttgart.de
Uni Stuttgart Tel.: 0711-7816-215
^ permalink raw reply [flat|nested] 10+ messages in thread
* Re: Optimization for PII
1999-09-13 13:00 ` Marc Lehmann
@ 1999-09-30 18:02 ` Marc Lehmann
0 siblings, 0 replies; 10+ messages in thread
From: Marc Lehmann @ 1999-09-30 18:02 UTC (permalink / raw)
To: gcc
On Mon, Sep 13, 1999 at 11:34:45AM +0300, Rainer Dorsch <rainer@rainer.informatik.uni-stuttgart.de> wrote:
>
> The egcs documentation says
>
> `-march=CPU TYPE'
> Generate instructions for the machine type CPU TYPE. The choices
> for CPU TYPE are: `i386', `i486', `pentium', and `pentiumpro'.
> Specifying `-march=CPU TYPE' implies `-mcpu=CPU TYPE'.
>
> It does not mention PII and PIII. Does pentiumpro run on PII and PIII or is
> there another undocumented option (I am using gcc-2.95.1)?
i386 == i486 == pentium == k6 < pentiumpro == pii == piii
(e.g. i386 code runs on pii, but pii code not on i386)
This is the view the compiler currently has. The pii and piii additional
instructions are not yet used, as isn't mmx.
It is unclear (to me) wether this will change, as there is no real strong
hierarchy anymore (chip can have mmx but not fcomi, or vice versa), or
wether additional switches (-mmx, -m3dnow) will be introduced.
--
-----==- |
----==-- _ |
---==---(_)__ __ ____ __ Marc Lehmann +--
--==---/ / _ \/ // /\ \/ / pcg@goof.com |e|
-=====/_/_//_/\_,_/ /_/\_\ XX11-RIPE --+
The choice of a GNU generation |
|
^ permalink raw reply [flat|nested] 10+ messages in thread
* Re: Optimization for PII
1999-09-13 21:34 ` Jeff Garzik
@ 1999-09-30 18:02 ` Jeff Garzik
0 siblings, 0 replies; 10+ messages in thread
From: Jeff Garzik @ 1999-09-30 18:02 UTC (permalink / raw)
To: Richard Henderson; +Cc: gcc
Richard Henderson wrote:
>
> On Mon, Sep 13, 1999 at 05:20:50PM -0700, Mike Stump wrote:
> > I think we could add PII and PIII as essentially aliases to pentiumpro
> > for users that don't know much about the CPU line.
>
> I thought about this. It starts to make the specs file extremely ugly.
> Just documenting things properly is probably good enough?
IMHO supporting -march=p2 and -march=p3 is a nice user friendliness
feature. It makes transitioning to P-II- and P-III-specific
optimizations easier, as folks won't have to change their Makefiles to
take advantage of new compiler features.
Regards,
Jeff
--
Custom driver development | Never worry about theory as long
Open source programming | as the machinery does what it's
| supposed to do. -- R. A. Heinlein
^ permalink raw reply [flat|nested] 10+ messages in thread
* Re: Optimization for PII
1999-09-13 17:21 Mike Stump
1999-09-13 21:02 ` Richard Henderson
@ 1999-09-30 18:02 ` Mike Stump
1 sibling, 0 replies; 10+ messages in thread
From: Mike Stump @ 1999-09-30 18:02 UTC (permalink / raw)
To: gcc, rainer.dorsch
Yes, pentiumpro can (should) be used on the PII and PIII. I think we
could add PII and PIII as essentially aliases to pentiumpro for users
that don't know much about the CPU line.
^ permalink raw reply [flat|nested] 10+ messages in thread
* Re: Optimization for PII
1999-09-13 21:02 ` Richard Henderson
1999-09-13 21:34 ` Jeff Garzik
@ 1999-09-30 18:02 ` Richard Henderson
1 sibling, 0 replies; 10+ messages in thread
From: Richard Henderson @ 1999-09-30 18:02 UTC (permalink / raw)
To: Mike Stump; +Cc: gcc, rainer.dorsch
On Mon, Sep 13, 1999 at 05:20:50PM -0700, Mike Stump wrote:
> I think we could add PII and PIII as essentially aliases to pentiumpro
> for users that don't know much about the CPU line.
I thought about this. It starts to make the specs file extremely ugly.
Just documenting things properly is probably good enough?
r~
^ permalink raw reply [flat|nested] 10+ messages in thread
* Re: Optimization for PII
1999-09-13 21:02 ` Richard Henderson
@ 1999-09-13 21:34 ` Jeff Garzik
1999-09-30 18:02 ` Jeff Garzik
1999-09-30 18:02 ` Richard Henderson
1 sibling, 1 reply; 10+ messages in thread
From: Jeff Garzik @ 1999-09-13 21:34 UTC (permalink / raw)
To: Richard Henderson; +Cc: gcc
Richard Henderson wrote:
>
> On Mon, Sep 13, 1999 at 05:20:50PM -0700, Mike Stump wrote:
> > I think we could add PII and PIII as essentially aliases to pentiumpro
> > for users that don't know much about the CPU line.
>
> I thought about this. It starts to make the specs file extremely ugly.
> Just documenting things properly is probably good enough?
IMHO supporting -march=p2 and -march=p3 is a nice user friendliness
feature. It makes transitioning to P-II- and P-III-specific
optimizations easier, as folks won't have to change their Makefiles to
take advantage of new compiler features.
Regards,
Jeff
--
Custom driver development | Never worry about theory as long
Open source programming | as the machinery does what it's
| supposed to do. -- R. A. Heinlein
^ permalink raw reply [flat|nested] 10+ messages in thread
* Re: Optimization for PII
1999-09-13 17:21 Mike Stump
@ 1999-09-13 21:02 ` Richard Henderson
1999-09-13 21:34 ` Jeff Garzik
1999-09-30 18:02 ` Richard Henderson
1999-09-30 18:02 ` Mike Stump
1 sibling, 2 replies; 10+ messages in thread
From: Richard Henderson @ 1999-09-13 21:02 UTC (permalink / raw)
To: Mike Stump; +Cc: gcc
On Mon, Sep 13, 1999 at 05:20:50PM -0700, Mike Stump wrote:
> I think we could add PII and PIII as essentially aliases to pentiumpro
> for users that don't know much about the CPU line.
I thought about this. It starts to make the specs file extremely ugly.
Just documenting things properly is probably good enough?
r~
^ permalink raw reply [flat|nested] 10+ messages in thread
* Re: Optimization for PII
@ 1999-09-13 17:21 Mike Stump
1999-09-13 21:02 ` Richard Henderson
1999-09-30 18:02 ` Mike Stump
0 siblings, 2 replies; 10+ messages in thread
From: Mike Stump @ 1999-09-13 17:21 UTC (permalink / raw)
To: gcc
Yes, pentiumpro can (should) be used on the PII and PIII. I think we
could add PII and PIII as essentially aliases to pentiumpro for users
that don't know much about the CPU line.
^ permalink raw reply [flat|nested] 10+ messages in thread
end of thread, other threads:[~1999-09-30 18:02 UTC | newest]
Thread overview: 10+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
1999-09-13 2:35 Optimization for PII Rainer Dorsch
1999-09-13 13:00 ` Marc Lehmann
1999-09-30 18:02 ` Marc Lehmann
1999-09-30 18:02 ` Rainer Dorsch
1999-09-13 17:21 Mike Stump
1999-09-13 21:02 ` Richard Henderson
1999-09-13 21:34 ` Jeff Garzik
1999-09-30 18:02 ` Jeff Garzik
1999-09-30 18:02 ` Richard Henderson
1999-09-30 18:02 ` Mike Stump
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).