If we compare the tools we had available to us 25-30 years ago to the tools we have
today, we would unimpeachably call that period the dark ages of silicon development. The steps
for developing a chip rearward then and now are such(prenominal) the same in a broad sense: definition,
logic and enlistment design, verification, layout and mask making, silicon wafer fabrication
(processing), and debug and test. But that is where the similarity ends. In the early days,
design, verification, and testing were done manually for the most part. Fortunately, the chip
designs back then contained fewerer than 30,000 transistors instead of todays 42 million.
Today, chip definitions direct specifications hundreds of pages long, logic design is largely
a matter of writing software code, computers step on it millions of verification tests on logic and
timing in a few days, and testing is done on multi-million dollar testers
When Peter fall in Intel in 1974, he was the sole design engineer on the 5810. The
product definition process for that chip illustrates a radical deflexion between the Intel of
then and the Intel of today. His boss, Joe Friedrich, prepared a single page document
called a scar Specification (spec) that gave the four-digit name to the product.
It also
gave the pinout and defined the function in sufficient detail for the approving parties to
decide whether they wanted to build it.
The wide-cut chain of command of Intel, from Robert Noyce on down to Joe Friedrich, met
in a style to decide whether to approve development based on the 5810 Target Spec. In
that single meeting, the decision was made to proceed. The product name, 5810, remained
constant from that point fore throughout the product life. The name appeared in the
Target Spec, the schematics, any memos, the true layout, the masks, the...
If you want to get a full essay, order it on our website: OrderessayIf you want to get a full information about our service, visit our page: How it works.
No comments:
Post a Comment