Best 1970s Microprocessor

[towforce]

The 8086 had a completely different architecture.

From link: "Whereas the 8086 was a 16-bit microprocessor, it used the same microarchitecture as Intel's 8-bit microprocessors (8008, 8080, and 8085). This allowed assembly language programs written in 8-bit to seamlessly migrate"

That's just not true, but more a part of Intel's marketing. Of course the 8086 would be very close to its immediate predecessor the 8085. You can see that in the names, right?

In practice the 8-bit assembly could not be directly reused in the 8086, but had to first be run through a converter program. And the resulting code was nothing like what new 8086 code would look like.

But as it had a 16-bit memory bus and ran at 5 MHz, the converted code was still faster than the original. So it "worked", sort of.

Very interesting and wholly believable!

If, before voting, you'd like to see these devices running, here's a web page that lets you run a program step by step (or automatically) and see exactly which parts of the CPU are being activated (it took dedication to build these web pages!).

6502: http://www.visual6502.org/JSSim/index.html

Z80: http://www.visual6502.org/JSSim/expert-z80.html

This gives you a view that the heroes of dedicated chess computers toiling away in assembly language probably didn't have: I suspect that for them, these devices were "black boxes".

[towforce]

Nice comparison of the performance of the Z80 versus the 6502 - link.

One opinion there: the 6502 is better for general purpose computers whereas the Z80 is better for embedded devices. I would argue that a dedicated chess computer is more like an embedded device than a general purpose computer.

[mclane]

I wonder why there is no stronger chess prg. for the ti99/4a.

There is one, but it is not very strong.
Maybe it was too early or because nobody was used to the cpu.

https://en.wikipedia.org/wiki/Texas_Instruments_TMS9900

[towforce]

I wonder why there is no stronger chess prg. for the ti99/4a.

There is one, but it is not very strong.
Maybe it was too early or because nobody was used to the cpu.

https://en.wikipedia.org/wiki/Texas_Instruments_TMS9900

Maybe a clue in the linked article: the 64 pin format was, apparently, difficult and expensive to work with at the time.

[Ras]

I wonder why there is no stronger chess prg. for the ti99/4a.

Because it was a misdesigned product - essentially a 16 bit system where the 8 bit bus dragged it down to 8 bit performance, only more expensive to produce. The 8 bit competitors started a price war, TI had to follow and tried to make money back by being the sole publisher for software - which crippled the software supply, including chess programs.

[Ras]

One opinion there: the 6502 is better for general purpose computers whereas the Z80 is better for embedded devices.

That was true only later, but not at the time. What was (and still is) used in embedded is not the Z80 processor, but Z80 based microcontrollers. The difference is additional on-chip periphery that pure microprocessors don't have, such as GPIO, ADC, SIO and the like. By the time Z80s were extended to that purpose, the cost advantage of the 6502 wasn't as large anymore, but the Z80 was easier to program, especially with the interrupt scheme better suited for embedded realtime stuff.

Back at the time, the Z80 was good for real computers, mostly because it could run CP/M and CP/M applications, which was more the business line. The 6502 on the other hand offered better bang for the buck and competed heavily on price. That made the 6502 better for home computers as well as chess computers that didn't even have any operating systems.

I would argue that a dedicated chess computer is more like an embedded device than a general purpose computer.

Dedicated chess computers often ran general purpose CPUs, see the H+G product lines not only with 6502, but also the more expensive ones with 68k / 020 / 030. Today, we do have microcontrollers that offer the performance as well as nicely intergated peripherals, e.g. ARM Cortex-M, but not back then. You had to choose between computing and integrated peripherals.

In the mid-70s, the first microcontrollers were available, such as TI TMS1000 or Intel 8048 (predecessor of the well-known 8051), geared towards use e.g. in embedded systems in cars - but they were rather simple chips with not much computing power and would not have been able to compete in chess computers. Their point was to simplify embedded designs that didn't need much computing power so that these chips would fit because they also eliminated external helper chips, given that periphery was on-chip already.

[towforce]

Thank you to everyone who has contributed knowledge and experience. I lived through the age when it became possible to have a computer at home and on one's desk, which the 8-bit microprocessors first made possible, and I have been impressed by how much new insight I have acquired just from this thread!

[towforce]

My 'matchbox computer' used 6502, and did not need anything more than the CPU, an 8K x 8 memory chip, an 8-bit latch as output port for the 7-segment LED display and one more TTL chip with gates for generating the control signals for that latch and memory, and the clock signal.

Thinking about it, your computer sounds similar to Sinclair's first computer - the MK14, launched in 1977 for £40 (link). This used the SC/MC 8-bit microprocessor, which, apart from this product, was not very successful (its other main use was as an elevator controller).

[Eelco de Groot]

My 'matchbox computer' used 6502, and did not need anything more than the CPU, an 8K x 8 memory chip, an 8-bit latch as output port for the 7-segment LED display and one more TTL chip with gates for generating the control signals for that latch and memory, and the clock signal.

Thinking about it, your computer sounds similar to Sinclair's first computer - the MK14, launched in 1977 for £40 (link). This used the SC/MC 8-bit microprocessor, which, apart from this product, was not very successful (its other main use was as an elevator controller).

A picture for those who have not seen it before, the original "Fire on Board" supercomputer, this is from the ICCA journal originally, photo Jos Uiterwijk. Detailed account from HGM also in Computerschaak, digitalized by Hein Veldhuis here in this pdf

Read also about the journalist who erased eproms in the very first Dutch computer chess championships, 1981 or 1982 and more on the matchbox chess computer on Adam Hair's chess blog http://adamsccpages.blogspot.com/2012/0 ... puter.html

[towforce]

A picture for those who have not seen it before, the original "Fire on Board" supercomputer, this is from the ICCA journal originally, photo Jos Uiterwijk. Detailed account from HGM also in Computerschaak, digitalized by Hein Veldhuis here in this pdf

The good news: Google can translate documents for you at https://translate.google.com/

The bad news: most of the document is images of text, not text itself, and doesn't get translated!

A couple of questions:

1. in which year was the matchbox computer built?

2. how are moves input?

3. How does the computer indicate its move?

Of course, if someone wrote a program for a smartwatch, it would thrash this program - but it's VERY impressive for its time!