[illumos-Developer] Ugh... xdr_{float,double} problems

[Deano]

Some off the top of my head are,

Cell's SPU floats are non IEEE 754, as are some vector ISA (ARM Neon,
PowerPC VMX for example), also some GPUs aren't (but that’s not a real
concern here :D)

Both Cell and platforms with non IEEE754 vector ISAs have units that are
IEEE 754 so conversion wouldn't be an issue.

Most modern ARM platforms ARE IEEE754 (though likely not the latest
revision) which is likely the main other platform to be worried about imho.

HTH,
Deano

-----Original Message-----
From: Jason King [mailto:jason.brian.king at gmail.com] 
Sent: 31 January 2011 15:54
To: Joerg Schilling
Cc: developer at lists.illumos.org
Subject: Re: [illumos-Developer] Ugh... xdr_{float,double} problems

This is why I asked :)  Are there actual identifiable platforms in use
today (that would also be a suitable target for Illumos -- some 50Mhz
CPU w/ 16mb of ram is unlikely to ever be relevant for Illumos :P)
that do not use IEEE 754?

Even then, if there are platforms, how reasonable is it to assume /
expect that they write code to convert to IEEE 754?  It seems like it
would be desirable since such conversion code could utilize knowledge
of the underlying implementation.

On Mon, Jan 31, 2011 at 8:59 AM, Joerg Schilling
<Joerg.Schilling at fokus.fraunhofer.de> wrote:
> Jason King <jason.brian.king at gmail.com> wrote:
>
>> Not a big deal.. one of the downsides of online communication is
>> sometimes things don't always come through that might in person..
>>
>> I'm tempted to just drop the code completely.  Are there any platforms
>> in widespread use today that don't use IEEE 754?  The code suggests
>> that as far back as 1985, vax was the only one that didn't.  If there
>> is, my guess is that there's probably going to be code specific for
>> that platform that's going to be more accurate and faster to translate
>> anyway.
>
> If you have a FP hardware in the system, you may be right.
>
> There are however embedded systems that may use a different implementation
for
> performance reasons - so I am not sue whether your assumption is correct.
>
> Jörg
>
> --
>  EMail:joerg at schily.isdn.cs.tu-berlin.de (home) Jörg Schilling D-13353
Berlin
>       js at cs.tu-berlin.de                (uni)
>       joerg.schilling at fokus.fraunhofer.de (work) Blog:
http://schily.blogspot.com/
>  URL:  http://cdrecord.berlios.de/private/ ftp://ftp.berlios.de/pub/schily
>

_______________________________________________
Developer mailing list
Developer at lists.illumos.org
http://lists.illumos.org/m/listinfo/developer