Re: ZIP 8Mhz Apple IIc vs 66MHz IBM Clone: Computer Race

[tm@netcom.com (Toshiyasu Morita)]

musjndx@gsusgi2.gsu.edu (Jonathan N. Deitch) writes:
 
> sebugg@presby.edu (Stephen Buggie) writes:
> 
> 
> >                       SOFTWARE SUGGESTIONS NEEDED:
> >      SPEED RACE BETWEEN  8 MHZ APPLE IIC AND  66 MHZ  486DZ ZEOS
> >      ===========================================================
> 
> 
> >A colleague at my college has used his new faculty grant to buy an 
> >IBM clone for his office: 17" color monitor, laser printer, and thi
> 
> >                ZEOS 486 DXZ  66 MHz,  8 meg RAM, with
> >                 Diamond Speed 5+ duo graphics card
> 
> >I bragged to him that my office IIc now has an 8 MHz ZIP CHIP and t
> >screens change almost instantly.  I read on this bbs too, that Appl
> >more data processing in a single cycle than do IBMs, so the "proces
> >speed" of the Apple II is underestimated.  I hinted to him that the
> >speed" of our computers may be equal.  Needless to say, he is skepi
> >my claim.  He is convinced that his computer will fly circles aroun
> >the biggest Macs.
> 
> Well ... this is true mostly.  Apples use a 6502 processor which,
> when  --MORE--(37%)
> 
> comparted to a 486 or Pentium is practically a RISC chip.  Most
> 6502 instructions take 3 or less cycles to complete.  But the
> disadvantage is you can't do much with a simple little 8bit
> processor.  Anything math intensive (ie: lots of cycles) the 486
> will clean up big.
 
Uhh...I hate to tell you this, but a 80486 executes most 
instructions in a single clock cycle, and the Pentium can 
dual-issue (execute two instructions simultaneously) most of the time. 
 
> Anything involving small (<64k) memory moves and the apple will
> probably win due to a lesser amount of time needed to make the move
> (smaller # of cycles even with a slower clock speed).  Note that
> the 486 may have a ram cache which will render this 6502 advantage
> nil.
 
Urgh, on the 80386/80486 there is a block memory move 
instruction which can move a longword (32-bits) in 7 clock 
cycles. The 6502 will take about 17 clocks per byte (lda (source),y sta 
(dest),y dey bne) and the system clock is slower.
 
> Also note that the 486 has a much bigger address space than a 6502
> and is used to working with larger data sets.
> 
> >I suggested that a "computer race" be arranged with my ZIP IIc vs h
> >486.  What SOFTWARE could be used that would be equivalent for the 
> >computers?  The software should have theser characteristics:  take 
> >amount of time to complete (20-60 sec),  not rely on excessive memo
> >only 128K in the IIc  vs. 8 megs in his), and not require a hard di
> >has a big one while the IIc has none.
> 
> A simple memory move benchmark should work okay.
> 
> >I have plenty of 8 bit Apple software; not sure what he has.  But s
> --MORE--(67%)
> 
> >from readers will be appreciated.    (Note:  If my IIc will be abso
> >whipped in any competition with the 486, advise me now so I could c
> >the "tortoise and hare" computer race!   B-)
> 
> I think the 486 will still win ... there's almost a 9x speed
> advantage in a 66mhz 486 vs a 8 mhz 6502.  Remember that any i/o
> the Apple does will have to be done in 1mhz mode as the Apple's
> slots run at that speed.  The 486 doesn't have this problem.
 
I program IBM PCs (in C & 386/486 assembly) and I've worked in 
6502/65816 assembly, and trust me, even a 386 is much faster than a 6502.
 
80386 runs most instruction in about 3 clocks - about the same 
as a 6502, plus it has 32-bit registers and can run at
33 or 40 Mhz vs. 8 Mhz...
 
I write a lot of graphics tools for the IBM PC, and I can 
write every single line in C (even the interrupt code) -
I don't think you could write a tolerable paint program in Aztec C
on an accelerated Apple II.
 
My lowly 386/40 at home benchmarks at 19,000+ Dhrystones, and I 
had a 486/33 at work that would do about 44,000+ Dhrystones 
(if I remember correctly).
 
I think I remember somebody posting about a GS benchmarking
600+ Dhrystones at 2.8 Mhz, so at 10 Mhz you'd get about 2200+ 
Dhrystones or so.
 
I'm as much a diehard Apple II fan as anybody (I've been programming
Apple IIs since '86 and I have three Apple IIs) but please...
comparing the speed of any Apple II to a 486/66 is like 
drag-racing a stock Model T against a Lamborghini Diablo.
(Model Ts are more fun than a Diablo, though)
 
(long .sigs deleted)
 
 
Toshi Morita - Manager of New Technology @ LucasArts Entertainment
tm@netcom.com