Re: Remember "Bit-Slice" Chips ?

On 12/7/24 9:29 AM, The Natural Philosopher wrote:

On 07/12/2024 07:33, 186282@ud0s4.net wrote:

Latter 70s they were The Thing.
>
Needed a 64/128 processor in an 8-bit world, then
bit-slice processors were yer fix.
>
They were the basics of a CPU - but wired so you
could physically attach them to MORE processors.
All the necessary flags/registers/etc could be
expanded wider and wider.
>
You could buy 2-bit, 4-bit, slice processors and
physically build something much stronger.
>
I even remember hearing of them mentioned in some
cheap TV series - some geek with his own R2D2
clone that was WAY too capable for the era.
>
TODAY ... well ... you can make a 64/128 on like
a 1cm die - really party on a 2cm die.
>
Bit-slice now - you'd loose far too much in
the interface wiring. Really no longer a
solution - unless maybe you need a 1024/2048
processor  :-)
>
Kinda the same goes for 'Transputers' - parallel
solution using ultra-speed (for the day) serial
links between many processors to coordinate
things between all the chips (they could have
a shared memory area too).
>
Older tech limitations spawned FIXES ... there
were many. Some were very *clever* - might even
have future apps.
>
>

As an engineer it always amazes me on how such little things the success of a technology depends.
 Aircraft could have been invented hundreds of years earlier if a lightweight power source had turned up, but steam wasn't good enough, and it took oil petrol or gas to make the power sources light enough.

   SOME of Da-Vinci's designs WOULD have worked if they'd
   had a chain-saw engine and prop. CONTROL ... well ...

electric cars fail because the batteries simple are barely good enough to replace IC.

   Barely ... sometimes less than that.

Windmills and solar panels are useless for the same reason. - there is no storage able to meet the intermittency problem.

   Well, Musk sells gigantic lithium batteries  :-)
   Some kinds of large fiber-based flywheels might
   do the trick - but they'd better be buried just
   in case ....

Yet the advance of photolithography and quantum theory made the integrated circuit a possibility, the Cold War mandated the need for small light electronics for missiles, and here we are.

   We get along, we make progress in various ways.
   But, at least now, I don't see the "next wave" -
   something like transistors or IC engines or such.
   They promise 'quantum' and such, but the MEANS
   work against it.
   The FET was visualized in the very early 1900s by
   some Russian prof. Digital processing not long
   after by another Russian (some proto's survived).
   The general-purpose computer was Babbage/Lovelace
   in the mid 1800s. The issue was a BETTER MEANS
   of realizing the CONCEPT. That took awhile.
   Every time that 'better means' arrived - HUGE leaps.
   Sometimes it was SMALL stuff - but provided critical
   leverage.