On Wed, 30 Jun 1999 11:46:13 +0200, Anders Haglund wrote:
>
>Matt Giwer wrote:
>> Jurica Lovakovic wrote:
>>
>> > Are there any plans for a P3-optimised compile of POV-Ray?
>>
>> > Or, even better, can POV-Ray make use of the new P3 instructions (I
>> > suppose it can, because they're supposed to accelerate 3d computations).
>>
>> That is game 3D. It has nothing to do with rendering 3D. They
>> increase the speed of manipulating pre-rendered objects in games.
>> Things like POV-Ray should work faster on RISC machines but I am
>> unaware of a comparison tests on them.
>The basic 3D operations are the same, like matris and vector operations.
>And if I  remember right the p3 have instructions just for those kind of
>operations. But to use them would require a rewrite of the matris and vector
>code used in povray, i't can't be done by just setting the compilator to
>'P3 optimize'. Anyway, to rewrite the code shouldn't be too hard but I don't
>think the speed increase would be very big.

Unfortunately, as with 3DNow! before it, the SSE instruction set is 
single-precision only.  Most of the math in POV is double-precision, 
and with good reason.


Chapter 9)

   9.1. OVERVIEW OF THE STREAMING SIMD EXTENSIONS

   The Streaming SIMD Extensions introduce new, general-purpose, floating-point 
   instructions, that operate on a new set of eight 128-bit SIMD floating-point 
   registers. This set enables the programmer to develop algorithms that can 
   finely-mix packed, single-precision, floating-point and integer using both 
   Streaming SIMD Extensions and MMX instructions respectively.

Post a reply to this message