On Thu, 30 Mar 2000, James McCartney wrote:
> on 3/29/00 7:49 PM, David Olofson at [EMAIL PROTECTED] wrote:
>
>
> > Yes, but it's still SIMD with special op-codes; not magic. How to make
> > an optimizer unroll loops into efficient code using packed data types
> > and SIMD instructions? I'm sure it's possible to do in theory, but I
> > seems to me that the complexity suggests AI solutions.
>
> Actually it only requires better languages without the pointer aliasing and
> side effect ambiguities that C has. Functional languages are better suited
> for automated parallelizing. The Sisal language, now defunct unfortunately,
> was an ideal language for this sort of thing. It was easily parallelized at
> whatever level of granularity most appropriate to the hardware and could
> outperform C and Fortran numeric intensive code.
Yes, but almost all code of this kind is still being written in C or
C++... The usual "generally accepted standard" vs. "new and possibly
better" conflict. Functional languages have been around for quite a
while without a real breakthrough. OTOH, perhaps the SIMD technology
will finally change this...
//David
.- M u C o S --------------------------------. .- David Olofson ------.
| A Free/Open Multimedia | | Audio Hacker |
| Plugin and Integration Standard | | Linux Advocate |
`------------> http://www.linuxdj.com/mucos -' | Open Source Advocate |
.- A u d i a l i t y ------------------------. | Singer |
| Rock Solid Low Latency Signal Processing | | Songwriter |
`---> http://www.angelfire.com/or/audiality -' `-> [EMAIL PROTECTED] -'
