historical/interest off-topic Apollo lunar mission cameras

Mike O'Dell mo at ccr.org
Tue Aug 28 21:39:42 CDT 2007

Robert E. Seastrom wrote:

> Should you want to build an Apollo Guidance Computer in your own home,
> you can find most of the information you need here:
> http://klabs.org/history/build_agc/

that's a remarkable amount of work. but if you're
going to completely re-engineer the gate level,
not sure why i wouldn't use a big FPGA to do it.

the really interesting thing about the original machines
is that they were asynchronous logic, not synchronous
logic as the replica. asynchronous logic is enjoying
a real renaissance these days, spearheaded by some
guys from Caltech who formed Fulcrum Microsystems
(fulcrummicro.com). Fulcrum did the first single-chip
10-gig ethernet switch and their is still the leader
in terms of port density and lowest power consumption.
the chip is completely asynchronous and has very
low delay as well. it's being used in a number of 10-gig
switches on the market now and is driving the cost
per port into the gig-E range pretty quickly.

when the company was first formed, they had this ultra-cool
design capability for fully-async logic so the question was
what to do with it. the first thing they designed
was a fully-async MIPS R20K which was blisteringly fast
but had a huge marketing problem - they couldn't quote
a "clock rate" for the part! and the sales prospects
for yet another processor part, no matter how good, were
pretty ugly. so they cast about a while trying to think
of another part to use as the flagship demo of their
design capability and thought of a 10-gig ethernet switch chip.
it was originally intended to demonstrate their abilities,
but the damn thing caught on like wildfire so now they are
a network silicon company. of course, they can still do
other stuff but the product focus is on the 10-gig ethernet

FFD: NEA was an early investor in Fulcrum and we remain
very active.


More information about the Tacos mailing list