6502

Otherwise know as a "supercomputer."

Grant Edwards wrote in news:4137d739$0$67058$ snipped-for-privacy@newsreader.visi.com:

Aside from the expected confusion of former Mitsubishi salesmen and FAE's not knowing anything about the Hitachi side (and I'm sure vice versa) early in the merger, we haven't had much problem with Renesas. One thing to keep in mind when dealing with them, though. Always go to the European website to look for information and tools. The US site doesn't have anything near the same amount of info.

That's definitely true. AFAICT, the Hitachi eval boards and a lot of support stuff all originally came out of Europe.

Well, looks like we have different ideas of the term "architecture". My definition is something like "The basic design of a system. How the components work together." In GB&FB's definition (not DEFINITION; there's no need to shout) two architectures can only be the same if the two products are the same. Would this also hold for the original meaning of architecture; that two buildings have the same architecture if they are exact copies? IIRC (CMIIW) the 6309 was backward object code compatible with and a pin-for-pin replacement for the 6809. How much closer can you get without being identical?

The discussion of removing one index register from the 6502 and its effect is interesting in that the 6520 was developed as an improved 6800. The second index register was one of the major improvements.

I know from experience of programming of a rather substantial DP application for the 6800 was a royal pain with only one index register. The programming model used simulated a second index register, resulting in a pseudo-architecture that had an amazing resemblance to a 6809 well before the

... snip ...

IIRC it was an independent architecture. MOS Technology (IIRC) developed it and the 6500 simultaneously. They withdrew the 6500, which was a 6800 clone, under threat of suits from Motorola.

And your credentials justifying your definition being more authoritative (or even just more appropriate to daily use) than that of Blauuw and Brooks would be...?

Even if we were to accept your criterion "how the components work together", the MC6809 and HD6309 would have distinct architectures, as the HD6309 has some components that are not present at all in the MC6809. The HD6309 components, therefore, must "work together" at least somewhat differently than those of the MC6809.

False. For example, Blaauw and Brooks consider the IBM 709 and IBM 7090 to have the same architecture. Yet the realization is much different; the 709 was built from tubes, and the 7090 from transistors. Perhaps more to the point, the IBM System/360 Model 40 and Model 65 have the same architecture, and are realized with the same family of MST circuits, yet their implementation (logic design) is almost entirely different.

Irrelevant; the word has a different meaning when applied to computers, and Blaauw and Brooks defined it in that context. A sensible person wouldn't try to describe the architecture of a computer using terms applicable to the architecture of a building. One doesn't hear about computers being of Greek, Neo-Classical, or Art Deco architectural styles. The use of the word "architecture" for computers is obviously by analogy, but one shouldn't try to push analogies too far.

Some people now use the word architecture to refer to the internal implementation details of the processor, such as how many pipeline stages are present and what microoperations a macrooperation is decomposed into; that is more properly termed "microarchitecture".

The IBM System/370 was essentially a superset of the System/360, yet the architects considered it to be a distinct architecture, because System/370 program would not necessarily run on a System/360. The HD6309 extends the MC6809 to an even greater extent than the System/370 extended the System/360.

If rather than claiming that the HD6309 and MC6809 share the same architecture, you instead claimed that the HD6309 architecture is an extension of the MC6809 architecture, I'd have no objection.

Or if you wanted to say that the MC6809 and HD6309 share the same "architectural style", I wouldn't argue. As far as I know, there has never been any effort to formalize a definition of architectural style as applied to computers. Would the MC6801 share that architectural style? What about the PDP-11? Hard to say. There are obvious similarities, but also significant differences.

That algorithm is insufficient when you are the semiconductor vendor and are attempting to determine what you should put in the chip in the first place. A much more thorough analysis is required.

My company had been dealing with both Hitachi and Mitsubishi, and we found that there was indeed a lot of confusion early on in the formation of Renesas, but that it seems to have settled down some. I would expect that the availability of both parts and information would be highly dependent on the local contact. In my experience (N. Carolina) it has been satisfactory, with the understanding that my expectations are tempered by long and sometimes bitter experience with parts vendors from around the globe. Caveat: my experience is mostly with getting tools and information. Usually others have dealt with actually getting parts and pricing.

Or the Japanese website, if you can understand what you read there.

This is true.

Ed