[SGVLUG] The tangential flogging continues (was: bit order (was Linux based web-server appliance))

[Dustin Laurence]

To give a concrete example of Mike's point (and mine, though I passed
over it pretty cursorially), this

http://www.picotux.com/

is a computer about the size of an Ethernet plug which is (as far as I
can tell) equal or a bit more capable than a 486(1) and it apparently
draws about 0.8 W.

Dustin

(1) OK, I'm guessing, because I'm not terribly familiar with the
architecture.  Let's do some useless research.  The PicoTux is a 32-bit
ARM running at 55 MHz--right in the ballpark of the later 486's (I think
they topped out at 100 or 133MHz, but 66 MHz is the one I remember as
being the common chip).  The real question of course is which does more
work per cycle.  OK....

http://www.digi.com/products/embeddedsolutions/ns7520specs.jsp

Tells us it is an ARM7TDMI family CPU.  The wikipedia article

http://en.wikipedia.org/wiki/ARM_architecture

doesn't list a 55 MHz example but shows it should do just about 1 MIPS
per MHz, so we'll estimate it does about 55 MIPS.  It would be nice to
have actual benchmark numbers rather than useless things like MIPS,
but....I didn't find a similar chart for a 486, but this

http://www.pcmech.com/show/processors/35/2

suggests the factor should be more like 0.8.  So let's say the PicoTUX
CPU is crudely equivalent to a 486DX/66, though since that's a
clock-doubled 33MHz chip I fear memory performance won't measure up.  In
any case, I have a strong suspicion that this is rather unfair to the
ARM architecture, but let it pass.

I wish I could remember how much memory 486's used to typically come
with.

-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 189 bytes
Desc: not available
Url : http://www.sgvlug.net/pipermail/sgvlug/attachments/20060525/47777e3c/attachment.bin