Did you try www.westerndesigncenter.com ? I just did it and it's as alive as ever.

I don't think they sell any 8MHz parts anymore though. Everything they have is faster now.

Garth

_________________
http://WilsonMinesCo.com/ lots of 6502 resources
The "second front page" is http://wilsonminesco.com/links.html .
What's an additional VIA among friends, anyhow?

They don't seem anxious to sell anything. I got as far with them as finding that they wanted to equivocate about the price/availability of their parts.

Applications demand high performance and high integration, these days, and it's not unreasonable to want a part with a full compliment of both ROM and RAM as well as all the desired peripherals on a single device. That's only achievable on programmable logic devices, since one can't anticipate what the desired peripherals might be. The WDC core is FAR too slow to be of interest, and it puzzles me that they want people to pay to use it. There are plenty of comparable 650x cores floating around as freeware.

Though it would be an interesting mental masturbation project to run a familiar software set on a really fast, say on the order of 20 MHz, CPU in an existing application, assuming it would run properly at all, or to build a platform that would run existing tools so one could produce more complex funcitons in firmware, there are too many processors out there by now that already perform that well to justify the effort if the bulk of the effort is in getting the chip supplier to sell you the chip.

Uli

I know you don't like to use the phone for pricing, but I had no problem when I just called up WDC, asked prices, and then said "give me five of this, five of that, and seven of the other," using my credit card. The whole transaction took a few minutes max, and the parts were at my door a couple of days later-- so I'd have to say it's not really a fair assessment to say that they don't want to sell anything. Keep in mind that their primary business is selling IP and they're not in the retail business, so we need to make sure our small orders are quick and easy for them to handle so they don't decide to quit selling hardware altogether.

I do wish Bill Mensch were more interested in selling a wider array of actual microcontrollers; but most microcontrollers on the market (regardless of manufacturer) don't have the memory to do what the part is otherwise capable of, and adding it externally means forfeiting a lot of the I/O that made the microcontroller so attractive in the first place. If I were to make a uC in an FPGA or something like that, I'd want to make sure I could get a few K of RAM on it and 32K or more of ROM.

My attraction to the 65k family is largely due to its high power-to-complexity ratio. Although there are other processors that deliver hundreds of MIPS, I don't think of the newer 6502 derivatives as being slow at all. The last automated test set-up I made (about 1992) used a 65c02 at only 2MHz on STD bus to control the signal generators, signal routing, programmable power supplies, signal conditioning, and 12-bit A/D convertors, plus the usual LCD, printer, keyboard, etc.. The system is capable of plenty of extra expansion beyond that. The processor speed was never a bottleneck in the actual operation of the ATE. The only thing I would like to have had more speed for was the compiling during the development of the 10,000 lines of code. That ATE is still in daily use now almost ten years later. I've used my 5MHz workbench computer for audio sampling experiments at 45,000 samples per second in the background while remaining interactive in the foreground. It has been fast enough for most of my work, but I'll still make the next one several times as fast. Since I can, I might as well. It will open up a few more possibilities while retaining the things that attract me to the 6502 (or in this case, the 65816).

_________________
http://WilsonMinesCo.com/ lots of 6502 resources
The "second front page" is http://wilsonminesco.com/links.html .
What's an additional VIA among friends, anyhow?

Actually, I don't like to use the phone for anything. I don't like the random interruptions, I don't like the always-inconvenient interrupt-driven nature of the thing. I much prefer the way email works. I send a message when I have a need to do so, and I see the reply when it's convenient to look. That way I don't have to give someone the "bum's rush" when I've got someone else here discussing something that might be affected by what he hears.

Moreover, it's MUCH more convenient to read what someone's written down, rather than listening to the hemming and hawing over something that hasn't been sufficiently thought out yet. It's bad enough wading through a disorganized, "stream-of-consciousness" written piece. Often with dreadful errors in syntax, grammar, and orthography, indicating that the fellow doing the writing doesn't see the effort as important enough to do it right.

If WDC's IP 650x core were competitive in performance, to the micrcocontroller cores already out there, it would be wonderful, but since it isn't ...

My problem with them has been their outright evasion of questions like "what do these things cost?"

I can't go to my clients, who still use 6502 applications I designed for them in the late '70's - early '80's timeframe, with the notion of upgrading their system with a software peripheral that will only run on a fast part if I can't tell them what it will cost. Estimated annual volume depends heavily on what it will cost, and I can't speculate. Moreover, I've no idea whatever what the WDC folks think their product is worth. I'd be really embarrassed if I told them it cost twice or three times what their current 4 MHz CMOS part (running at 5, BTW) when it's really 20 or 30 times that. I've asked for budgetary pricing on more than one occasion, only to get no reply at all.

As a consequence, I now have a couple of clients using current-generation 80C5x-core based products also of my design, that would have worked fine with their previous hardware and a faster 65C02. The 80C5x core does have instructions that allow for fairly fast table lookups, in fact faster than the 6502 at the same cycle rate but all bets are off if the 65C02 is running at, say, 15-20 MHz.

CPLD's and FPGA's, as well as ASIC's are large enough and fast enough that one can build a CPU into one with a substantial internal RAM or, alternatively a port to external RAM, that one should be able to run a 2.5 ns cycle given fast enough external RAM. The crux is the ALU design. A 650x core is simple enough that it can be small, and, with an efficient and efficiently used ALU doesn't have to consume a lot of logic. Since the stand-alone CPU can take as many as 8 (?) cycles to execute a single instruction, depending on core version and addressing mode as well as page boundary associations, even a very fast one takes quite some time to do the slowest things it does. It's probable that one implementing a core in hardware would modify his design to facilitate the operations that will most impact his task. That's an option you don't have with the fixed CPU design.

My own interests don't run as far as the 65816, since it's single-sourced. Complete Pentium or super-x86 board sets on PC-104 format boards are so competitive that it's hard not to use them. Software for almost anything already exists, so unless the CPU core is very fast indeed, it's likely I'd be inclined to buy a ready-made solution.

The bit with WDC has been that I'm not only P*SSED because my questions have gone unanswered and I've, as a result, lost competitive opportunities, but I'm disappointed at not being able to extend the life of my old designs when they're so inherently capable of surviving another decade, and simply because they, WDC, have been unhelpful.

Uli