🚀 go-pugleaf

ma106071@cs.brown.edu (Jim Lowe) writes:

>In article <BzF5MF.2Lv@usenet.ucs.indiana.edu>, ntaib@silver.ucs.indiana.edu (Iskandar Taib) writes:
>|> In article <1gdipoINN749@rave.larc.nasa.gov> w.l.vaughn@LaRC.NASA.GOV (Wallace L. Vaughn) write
> 
>|> >Mac II, 16mHz, system 7.01+, 8bit 640 x 480 video, 8 meg ram
>|> >Mac IIci, 25mHz, see above
>|> >386dx, 33mhz system, windows, 3.0, 640x480 svga screen, no cache. 8meg RAM
>|> >                     calc.(sec)       scroll(min:sec)
>|> >Mac II                    35.3            2:16.2 
>|> >Mac IIci                  15.3            1:31.2
>|> >Mac IIci/cache off        19.2            1:44.7
>|> >386dx                     22.5            2:25.3
>|> 
>|> Hmmmm.
>|> 
>|> Most 386 motherboards above 25 MHz have a built-in SRAM cache 
>|> these days. Even the bargain basement jobs.


>I'd say Motorola's 680x0 series is faster than Intel's 80x86 of the same
>generation, AT THE SAME CLOCK SPEED. The reason Intel's tend to be substantially
>faster in reality is that Intel always offers them at a higher clock speed than
>Motorola has yet achieved. e.g., the 33MHz 486 has been on the market for 2 years
>or so, while the 33MHz 68040, I believe, just came out within the last two
>months. (at least, the first Quadra based on it was released then).

But wasn't the 68040-33 clock doubled?  (i.e. didn't it run internally
at 66 MHz) Which would mean that the 68040-33 should be compared to the
80486DX2-66 . Which means that the Intel processor is still faster at the
same clock speed.

>Running a fast chip without a cache cripples it with so many wait states that it
>loses its clock speed advantage entirely. The original XT didn't need a cache
>because it ran at 4.77 MHz and used 210 ns DRAMs. (at 4.77 million clocks/sec.,
>each clock lasts exactly 210 ns).  Think about it; clock speeds have gone up 10
>times, while memory speeds have barely gone up 3 times. Trying to compare two
>fast chips, one of which has a cache, and one of which doesn't, is completely
>meaningless. Of COURSE the one with the cache will win by an enormous margin. As
>is pointed out above, for months or years now, virtually all 386 systems have
>been offered with cache.

And virtually all 386 systems are offered with accelerated video. Of course,
you could compare _one_ 386 system with a cached system - the system
with 33ns SRAM (and if I remember correctly, 4MB - anyone put a finger
on the thing? If I remember correctly, it appeared in a PC Magazine review)
:-)


>-- 
>*********************************************************************
>   ----- ----- --    --  |  -----
>     |     |   | \  / |  |  |
>     |     |   |  -   |     -----
>     |     |   |      |         |
>   --     ---  -      -     ----- 
>     GRANDIOSE
>		and
>			ANNOYING
>	.signature file.
>	jim lowe, cs141046@cs.brown.edu
>	          ma106071@cs.brown.edu
>**********************************************************************
>     
-- 
Ravikumar Venkateswar
rvenkate@uiuc.edu

A pun is a no' blessed form of whit.