Subject: Re: [linux-audio-dev] cpu/disk tradeoff
From: David Olofson (audiality_AT_swipnet.se)
Date: pe loka 15 1999 - 17:13:47 EDT
On Fri, 15 Oct 1999, Paul Barton-Davis wrote:
> >Well, it looks like x86 CPUs aren't very cost effective these days,
>
> just to clarify: the MIPS was about 5 times the speed *for this
> benchmark*, which tested a specific thing: table-driven sine wave
> generation. there is no particular reason to expect the difference to
> apply to everything else. it would be interesting to see the
> comparison for float computed sine waves.
Yes, I'm aware of that... RISCs are usually not very much faster than
CISCs in the *average* case, but DSP is far from the "average" code
run on workstations. RISCs usually do things they were not originally
designed for a lot better than CISCs, as CISC instructions turn into
legacy crap in such cases. (I could dig out some practical 68k
examples from my old Amiga code... And that's a *good* instruction
set, as opposed to the x86 mess.)
Besides, I'm still more interested in the Alpha than the MIPS (perhaps
I'm just missinformed?), but we'll see if/when the Athlon gets rid
of the 3DNow! waitstates flaw. (Hmmm... BTW, I saw some company
selling Athlon 500 nearly as cheap as Celeron 500. But what about
dual Athlon?)
//David
·A·U·D·I·A·L·I·T·Y· P r o f e s s i o n a l L i n u x A u d i o
- - ------------------------------------------------------------- - -
·Rock Solid David Olofson:
·Low Latency www.angelfire.com/or/audiality ·Audio Hacker
·Plug-Ins audiality_AT_swipnet.se ·Linux Advocate
·Open Source ·Singer/Composer
This archive was generated by hypermail 2b28 : pe maalis 10 2000 - 07:27:59 EST