On Thursday, April 21, 2022 at 8:39:05 PM UTC-4, Bob F wrote:
> On 4/21/2022 7:36 AM, Martin Brown wrote:
> > On 21/04/2022 14:44, rbowman wrote:
> >> On 04/20/2022 09:31 PM, Jasen Betts wrote:
> >>> On 2022-04-18, rbowman <bow...@montana.com> wrote:
> >>>>
> >>>> That's not a general problem. There was a period with the early Athlons
> >>>> that didn't implement some of the new Intel instructions but I've
> >>>> leaned
> >>>> towards AMD with no problem.
> >>>>
> >>>> It wasn't AMD but I recall one processor that ran CP/M and DOS, both
> >>>> rather poorly. National maybe?
> >>>
> >>> NEC V20
> >>>
> >>> https://en.wikipedia.org/wiki/NEC_V20
> >>>
> >> Right, National was mostly analog. I was surprised to find TI bought
> >> National. I hadn't thought about either in a long time. I worked on a
> >
> > National did have its not quite a VAX on a chip NS32k series.
> > Dec had its MicroVax 78032 (ISTR slightly later but better).
> >
> > Zilog had its Z8000. I had the tricky job of unpicking a way out of that
> > particular blind alley as one of my first jobs in industry. Olivetti M20
> > and certain of the dedicated word processors used it.
> >
> >> project that used the TMS9900. It's claim to fame was availability in
> >> a rad hard version. NEC seems to be rolling along.
> >
> > I worked on several using its smaller 9995 and bigger brother the 99k.
> >
> > I didn't appreciate at the time just how good at interrupt handling and
> > context switching it was until later when we tried to do the same real
> > time tasks on a notionally faster Motorola 68k series CPU.
> >
> The 8080 was so much faster than the 8008. I designed and wire-wrapped
> my first computers 8080 board to replace the 8008 board it started with.
> It handled the paper tape I/O much better. All the original parts came
> out of a dumpster at work.

Did it run the same code? I don't remember if you had to reassemble to go from the 8008 to the 8080 or not. I have an 8008 computer. It has not been fired up in years. The serial port was 20 mA current, 110 baud. I tweaked it to work with RS-232. Still, most async serial ports are TTL through a USB cable these days. But they work fine. 110 baud is hard coded into the ROM, yes, ROM. I think the unit was designed to use 1702A EEPROMs and even has a programmer, but I don't know if I've ever tired it.

The 8008 is a very slow CPU. It is the basis of the machine cycles on the 8080. The 14 bit address was multiplexed over the 8 bit data/address bus in two clocks and the data on a third clock, so T1, T2 and T3. Sound familiar? Even though the 8080 is not constrained by the 8 bit port (so the 8008 would fit an 18 pin package), it uses the same internal timing. Just at a much high clock rate. I think the 8008 I have is clocked at maybe 400 kHz.

--

Rick C.

--+- Get 1,000 miles of free Supercharging
--+- Tesla referral code - https://ts.la/richard11209