[vcf-midatlantic] Working on a historical microprocessor exhibt

[Dave McGuire]

On 2/11/20 11:11 AM, David Riley via vcf-midatlantic wrote:
>>  Well bit-slice chips were made by many manufacturers.  Keep in mind
>> the minicomputer manufacturers were the chip companies' bread & butter
>> for a long time.  Even Intel; they made the 3000 series of bit-slice parts.
>>
>>  But it wasn't really separate ICs that eventually merged to form
>> microprocessors.  The Am2901 was introduced in 1975, well after there
>> were a few microprocessors on the market.  These chips, and their
>> applications, coexisted with monolithic VLSI microprocessors for a long
>> time.  Bit-slice design was commonplace clear up until at least 1990.
>> It was just a different way of doing things that was more scalable and
>> flexible.
>>
>>  And at the time of introduction of the VAX-11/730, for example,
>> designing a single VLSI chip that implemented an architecture as complex
>> as VAX wasn't really practical.  The first single-chip VLSI
>> implementation of the VAX architecture (the 78032, used in the
>> MicroVAX-II and others) wouldn't tape out for another three years after
>> that.
> 
> And it was sloooow by comparison to its contemporary VAXen.  Its primary benefits were board space, component cost and power savings, at the expense of a fair amount of speed.  Fully integrated microprocessors didn't outperform their multi-chip or discrete-part brethren until the era of the NVAX (which was concurrent with the big-ass ECL-based VAX 9000, which it outperformed, much to DEC's chagrin).  That was 1989 or so.

  The 78032?  Oh heck yes it was slower than the "big boys".  But it got
an entire VAX processor, with FPU and 1MB of RAM, on a single board for
a fraction of the price of any of the "big boys".  It enabled the
deskside and desktop VAXen, etc.  Remember, there's more to life than
just speed! ;)

             -Dave

-- 
Dave McGuire, AK4HZ
New Kensington, PA