Hi --
--- In Peter Gammie <> wrote:
>
> On 03/08/2010, at 7:01 PM, wrote:
>
> > Is this buggy FPU only in the 200MHz versions of the ARM9-based TS-boards,
> > or does it also affect the 250MHz and 500MHz?
>
> It is in the ep93xx chips. You'll need to check what the other boards have on
> them.
>
> > Can I disable FPU usage all together in the kernel and/or clib so I'm
> > certain the FPU won't be used?
>
> I wouldn't worry about it, nothing uses the FPU by accident, and few can use
> it on purpose...
You can disable the Maverick Crunch support when you compile a 2.6 kernel, and
although enabling it later in userspace is possible, that is both hard (there
is an unlocking R/W sequence needed) and not recommended (there would be no
kernel support for preserving FPU state over context switches). And IIRC, you
have to use compiler flags to get gcc to even emit crunch instructions.
FWIW, I had no problems other than some trig precision when I used earlier
versions of Martins crunch patches. I didn't really pursue it though as moving
to EABI gave me all the FP improvement I needed (about 18-20 times speedup over
OABI). Crunch usage was a relatively small improvement compared to the
improvement going to EABI.
regards, ............ Charlie
>
> cheers
> peter
>
> --
> http://peteg.org/
>
------------------------------------
Yahoo! Groups Links
<*> To visit your group on the web, go to:
http://groups.yahoo.com/group/ts-7000/
<*> Your email settings:
Individual Email | Traditional
<*> To change settings online go to:
http://groups.yahoo.com/group/ts-7000/join
(Yahoo! ID required)
<*> To change settings via email:
<*> To unsubscribe from this group, send an email to:
<*> Your use of Yahoo! Groups is subject to:
http://docs.yahoo.com/info/terms/
|