1. Home
  2. Electronics Design
  3. How are serial digital sequences generated in code?
  4. Page 2

How are serial digital sequences generated in code?

I would put it the other way. The 8080 was very primitive, in fact one could consider it to be one of the first RISC processors :-). There was very little you without a large number of register to register operations. Look at the 8080 op codes (in octal), the first octal digit (two most significant bits) contained the main part of the op code, the next two octal digits (groups of three bits) contained either a register designation or selected between op codes.

The 6800 resembles somewhat Data General Nova. Although 6800 only had only two accumulators and an index registers, there were several addressing modes and some operations could be performed directly into a memory location, hence greatly reducing the need for a lot of registers.

An interesting side note that before the full technical specifications were published, much of the marketing material stated that the 8086 is compatible with the 8080. This might be true for some 8080 assembly source code which could be reassembled for 8086 with small modifications. However, when the binary op code specification was released, it became quite clear that there was no binary compatibility and the compatibility claim was quickly dropped from the marketing material.

At least the mechanics and peripherals were better than on Sinclair QL (68008 with 8 bit bus).

Reply to
upsidedown
Loading thread data ...

The 8086 and 8088 /were/ primitive, and they were considered old-fashioned and limited architectures when they were first produced. They did not compete with the 68k on performance, architecture, expandability, or planned future (the 68000 was designed to be the first in a family, the 8086/8088 designers didn't think beyond the end of next week).

The only think the 8086/8088 competed with the 68k on was better sales people, and lower system price with an 8-bit bus rather than a 16-bit bus. This combined with a half-wit pencil-pusher at IBM who felt that the IBM PC was merely a marketing exercise and no more than 1000 machines would be made - therefore, the technical aspects didn't matter, the machine had no future, and it was a good idea to override all the engineers (who wanted a 68k chip) in order to save a bit of cost. So instead of getting a well-designed, forward-looking 68k architecture we ended up with the x86 architecture, and we use PC's run by highly polished turds.

Reply to
David Brown

6502 wasn't exactly primitive either. Minimalist yes and it was working with that chip at Acorn that in part led to their ARM risc architecture.

formatting link

4004, 1802 and SC/MP were *truly* primitive.

It is a bit unfair to write the other bit players out of history. TI had the venerable TI9900 and 8 bit 9995 and a range of home computers. I didn't fully appreciate at the time how good the 99k was at realtime context switching until we tried to implement the same thing on a 68k.

Zilog had the Z8000 which Olivetti used in their ill fated M20. DEC had PDP11 on a chip (and later VAX on a chip) and there were sundry other workstation chips of various flavours around back then.

I can recall being in meetings where we would haggle about the number of

32k blocks of memory to add into a PC for shipment with instruments.

One the bad calls I made was in favour of adopting the NEC7220 graphics chip as used in Japan capable of up to 1024x1024 by 16 colours (or if very rich 256 with a daughter board). It could do two pages at 640x480 comfortably with hardware assisted drawing and filling.

--
Regards, 
Martin Brown
Reply to
Martin Brown

Besides, the AT was 6 MHz for the standard version and a blistering 8 MHz for the one with the racing stripes. ;)

Cheers

Phil Hobbs

--
Dr Philip C D Hobbs 
Principal Consultant 
ElectroOptical Innovations LLC 
Optics, Electro-optics, Photonics, Analog Electronics 

160 North State Road #203 
Briarcliff Manor NY 10510 

hobbs at electrooptical dot net 
http://electrooptical.net
Reply to
Phil Hobbs

6800/6502 were quite close to the DG Nova instruction set. While I haven't checked the instruction bit encoding, I would have assumed that some of the Nova instruction bits controlled directly the 74181 ALU chip :-). Anyway an easy way to implement microcoding and hence avoid the main memory instruction fetch bottleneck.
4004 and TMS1000 where just glorified four banger (BCD) calculators as derived from e.g. Texas TMS1802 calculator chip.

RCA CD1802 was a simple 8 bit processor with a lot of fast on chip registers.

The TI9900 architecture is idiotic (and TI9995 even worse) unless you have at least 64-256 words of on-chip cache.

DEC LSI-11/2 required (and the Ruskie copy the E60), required a whole dual/quad width Q-bus card. The F-11 (PDP-11/23) processor was on a dual width Q-bus card, the J-11 processor (PDP-11/73) was something close to a single chip processor. uVax-I/II was about a card size.

Reply to
upsidedown

On Thu, 15 May 2014 13:23:11 -0400, Phil Hobbs Gave us:

The ATA bus, however, was locked at 4MHz.

That is why the EISA machines *could* have been great, but were not.

My 25MHz 486 EISA would have been a screamer, if that bus was not locked.

PCI machines included that bus tertiary to PCI, as all busses were at that time(PCI years).

Try to find a 16 bit ATA slot-to-PCI or PCI-e. Probably out there. A lot easier to just fire up an old machine.

Reply to
DecadentLinuxUserNumeroUno

Am 15.05.2014 10:33, schrieb David Brown:

Bla... Intel could deliver volume, Moto had samples.

6 dB more transistors had its price.

cheers, Gerhard

(sorry for email)

Reply to
Gerhard Hoffmann

down like

like SATA 3

data

a feral

"dumbing

with

the

Core in 1970 has 110 ns access time and 370 ns cycle time; it just cost 5 times as much and used 20 times the power. More bigger faster minicomputers were bit-slice microcode architecture to 32 bits already. Remember the 2901 and 2911 chips? 4 k-words was a common small core memory size even then. 16 k-words and 32 k-words were ruling the roost.

?-)

Reply to
josephkk

At least in the embedded market for 8 bitters, Intel had a decent HLL compiler (PL/M-80) for 8080/8085), while most 6800 programming was done in assembler.

Later on, transitioning from 8080/8085 to 8086/8088 was easier than transitioning from say 6800 to 68000.

Reply to
upsidedown

When the pencil-pusher wanted max 1000 pieces, the price per processor and low availability should have been even less of a concern.

(No problem - the old "reply" button vs. "followup" button mixup, I assume.)

Reply to
David Brown

Yes, but assembly programming on the 68k is easy and fun, unlike assembly on the 8086.

Reply to
David Brown

ElectronDepot website is not affiliated with any of the manufacturers or service providers discussed here. All logos and trade names are the property of their respective owners.