Microcode .vs. Hardware (Was: hardware multiply/divide functionality in CPUs)

Christian Corti cc at informatik.uni-stuttgart.de
Thu Mar 10 04:26:50 CST 2011


On Wed, 9 Mar 2011, dwight elvey wrote:
> I never had any external storage. I can recall if it was tape of punch
> cards.

There was the optional Diehl Dilector
(http://computermuseum.informatik.uni-stuttgart.de/dev/decitrn/combi_s.html)

> It had two delay lines, one for data and the other for program.

That is what I thought at the beginning, too. But in reality one is the R 
delay line (the shorter one that holds 218 bits at 1MHz, plus R and Q 
flip-flops for a total of 220 bits) and the other is the M delay line 
(10889 bits plus M flip-flop). The R line holds two double-words 
(interleaved P and P_ bits for one I and one I_ phase (duration 110µs)), 
the M line holds 99 such words. Execution is from the R line that (in most 
cases but not always) also holds the data to be processed.

> I never did figure out how it coordinated the delay line with
> the tape. The tape didn't seem to have any speed control.

That's right, the boot-up process used a very tricky state-machine using 
the CPU flip-flops for holding the state and data from the tape. That's 
purely Frankel who likes using parts for many different purposes during 
different states of the machine. The loading phase is described in detail 
in the technical manual (I only have the German version).

Christian


More information about the cctalk mailing list