14 Jul
2015
14 Jul
'15
7:42 p.m.
On 7/14/2015 7:36 PM, Jon Elson wrote:
On 07/14/2015 07:44 PM, William Donzelli wrote:
On the system 360 CPUs, they did not use flip-flops like we are used to,
today. They used latches, making it a requirement that there be at
least two clock phases in most of the CPU, so that data into the ALU,
for instance, remained stable when some register at the output was
clocked. Since these were discrete transistor implementations, a real
flip-flop was too expensive, but a latch could be implemented in about 6
transistors, I think.
I guessing ( no schematic handy) that they made the 360 register file
easy to decode and build with latches.
IIRC, the
KB11 processors used in the DEC 11/45 and 11/70 (and other
related systems) used five "clocks delayed from each other" (more
commonly known as clock phases).
IBM used this method as well on many of their
machines.
The 11/45 used TTL ICs, so real FFs were available in
that technology,
although they may have used latches as well.
I have seen some 11/?? schematics on bitsavers that the alu uses AOI gates
and includes a latch term.
Jon
Ben.