Re: linux-2.3.31: drivers/char/drm/drmP.h does not allow 386 build

David Weinehall (tao@acc.umu.se)
Sat, 11 Dec 1999 01:40:37 +0100 (MET)


On Sat, 11 Dec 1999, Jamie Lokier wrote:

> David Weinehall wrote:
> > CONFIG_M386_GENERIC, which is the crufted version with compability for all
> > processors
> >
> > and
> >
> > CONFIG_M386, which is cruftfree and known to work on 386's, without any
> > optimisations for other processors.
> >
> > In my opinion, Linux is getting a bit too slow on 386's, without any real
> > reason.
>
> The related are for improvment is "optimise for 686 but compatible with
> all x86". Probably just a matter of setting the right compiler options
> and using the appropriate string and checksum routines.

Yes, but almost always when you perform "optimise for xxx but maintain
compability", you slow down (and/or increase memory comsumption) for all
the other alternatives that you don't compile for.

Tests performed by someone on the debian-devel list recently showed that
programs optimised for pentiums, were slower on _all_ other
processor-models (486/PPro/AMD etc.) than 386-code.

And the cache-aligning used in the kernel right now is at least too big
for both 386's and 486's.

It's not that I oppose optimisations, it's just that I want there to be
an 386-optimized version too; there is none now, only a "works for 386,
but probably butt-slow" option.

/David Weinehall
_ _
// David Weinehall <tao@acc.umu.se> /> Northern lights wander \\
// Project MCA Linux hacker // Dance across the winter sky //
\> http://www.acc.umu.se/~tao/ </ Full colour fire </

-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu
Please read the FAQ at http://www.tux.org/lkml/