[vorbis-dev] Vorbis bitstream specification...

[Segher Boessenkool]

Monty wrote:
> 
> On Mon, Feb 11, 2002 at 09:40:52PM +0100, Segher Boessenkool wrote:
> 
> > ARM works.  PowerPC works.  Both on multiple compilers.  Integer multiply
> > on x86 is a lost cause, so I never bothered to check.
> 
> That's just a *little* bit important.

Not really.  x86 can use floating point.  Widening multiplies are extremely
expensive on all current x86 cpu's.  Of course the C code _will_ work fine
(albeit slow), so you can still use it for testing.

> If you're very worried, you can always define it to be an assembler define
> like this (for ARM):

[snip]

> actually, "=&r"(dum) is needed to make that work (%1 may not overlap
> with %2 or %3, and GCC *will* try to do it), and "%r"(x) might be a
> good idea too.

I said *like* this :)  I'm happy enough it even parses.

I don't like gcc to switch x and y around; the number of cycles an ARM
multiply takes is dependent on the value of only one of the operands.  I like
to choose which one.

Cheers,

Segher

<p>--- >8 ----
List archives:  http://www.xiph.org/archives/
Ogg project homepage: http://www.xiph.org/ogg/
To unsubscribe from this list, send a message to 'vorbis-dev-request at xiph.org'
containing only the word 'unsubscribe' in the body.  No subject is needed.
Unsubscribe messages sent to the list will be ignored/filtered.