🚀 go-pugleaf

In article <24948@amdcad.AMD.COM> tc@diablo.AMD.COM (Tom Crawford) writes:
>In article <1614@astroatc.UUCP> ttl@astroatc.UUCP (Tony Laundrie) writes:
>>
>>Does anybody make a (256k 1Meg 4Meg) DRAM with ECL I/O?
>
>The short answer is "no, nobody makes dynamic RAMs of any size
>with ECL I/O".  There are, however, static RAMs with ECL I/O
>up to 256K.  Well, at least they are listed in the new IC Master.
>Try National NM5100.  National Hmmm.  Sounds like a Fairchild
>part to me.  Well, give it a try anyway.

It was indeed done at Fairchild, and it's an *excellent* design.  Spawned
quite a few copycat competitors, too -- some from the Far East and some
from (blush) the usa.  Rumor is that the Cray-3 uses >30,000 of these
for main memory.  (that wasn't a typo: 3x10^4).  Dunno what Cray pays
but I bet it's less than the quantity-10 price quote you might receive
over the phone :-).

Look at the digests of the chip design conferences; you'll find that
the Japanese chips use exactly the same ECL-to-CMOS merged decoder/level
shifter that Fairchild did -- only Fairchild published it two years
previously and the Japanese authors "forgot" to reference the prior
paper.
-- 
 -- Mark Johnson	
 	MIPS Computer Systems, 930 E. Arques, Sunnyvale, CA 94086
	...!decwrl!mips!mark	(408) 991-0208