Another whacked-out question that came to me in the shower this morining…
If Intel had continued with numbering its processors instead of naming them, where would we be now?
In the series that led through the IBM PC, there were the 8080, 8088, 8086, 80186, 80286, 80386, then the 80486. Then there was the Pentium. Then the Pentium Pro, Pentium II, Pentium III, Pentium 4. Then the Core and Core 2, some (most?) of which were Duo.
Which of these, if any, correspond to successive versions in the way that the 386 succeeded the 286? Is my laptop running on an 80986?
IIRC, the Pentium, Pentium 2, and Pentium 3 were all based on the x86 architecture, so they would’ve been 586, 686 and 786. The Pentium 4 is a different architecture, so it would not have had an x86 name.
I thought every later version was a superset of the earlier versions? Can’t my Core 2 Duo run 8088 code, even if in some byway away from the main action?
Different architectures, not instruction sets. Bear in mind that machine code is a fairly high-level language to a chip designer, and that there’s a lot of room for changes that don’t have any effect on software.
The Pentium is effectively the 586. Intel abandoned the numerical naming structure to differntiate its products from other manufacturers (Cyrix, AMD, etc) because their 586 models were little more then 486 revisions. (AMD’s fifth generation CPU was the K5, not it’s Am5x86, and Cyrix’s fifth gen CPU was it’s 6x86… but lets stay with Intel for now.)
Intel released the first version of it’s sixth generation architecture as the Pentium Pro. That core was revised multiple times, as the Pentium II, the Pentium III, and the Celeron. The Pentium 4 used a brand new architecture (NetBurst), and is considered the seventh generation CPU.
The Core architecture could be considered the eighth generation, but it also marks a departure from the standard progression in chip design (focus on multiple cores as opposed to increased clockspeed). The Core 2 is a revision of the original Core architecture. AMD and Intel are no longer in sync with each other, so it is hard to compare their current chips in terms of generational architecture.
I believe one prominent reason for the dropping of the numbers in favor of “Pentium” was that somebody in marketing figured out you can’t trademark a number.
I’m sure I wasn’t the only one hoping for the Pentium’s successor to be the “Sexium”.
Back in the early years of the Macintosh, the company used the letter “x” to indicate a computer with the Motorola 68030 chip (the first Mac had a 68000 chip). So, a Mac II with the 68030 was a IIx, a IIc became a IIcx, and so on.
Unfortunately, when they put a 68030 in a Mac SE, they called it the “SE/30”.
I actually tried calling mine an SEx now and then, but nobody got the joke. It’s SWIM chip eventually died, so now it has a mini-ITX motherboard in it. I call it an SE/86.
BTW: I don’t know from Intel, but I once knew more than I wanted to about AMD’s K7. As I recall, the core was a bunch of RISC processors sitting behind a massive preprocessor that sucked in x86 instructions in cache-line sized chunks and converted it into RISC code. Not really x86 at all.
Notice, BTW, that AMD droped the K prefix (at least externally) two generations before “K9.” Smart?
Yes, it is. That’s the defining feature: They boot into 16-bit Real Mode which is binary compatible with the 8086 and the 8088. (It is possible to tell whether you are running on an 8088 or not by using various tricks, but that has no bearing.)
The internal architecture doesn’t matter. The only thing that matters is what software it can run without the help of other software.
PS: In the real world, all modern 32-bit Pentiums are called i686. The only chips to be called something else are AMD-64 chips running in 64-bit mode. However, they’re still x86 because they still boot into 16-bit Real Mode and can still be put into 32-bit mode to run 32-bit x86 code.
Just to add a twist, the current “Core” stuff is a child of the Pentium III architecture. When the Pentium 4 came out, the previous generation was relegated to step-child status as a low-power laptop chip (Pentium M and variants). The P4 architecture came to an overweight, power-soaked end when the PIII/M iterations eventually surpassed it in speed with lower clocks and lower power, to become the Core series. No cite, but I’m friendly with some folks at Intel because of my work, and it’s also fairly widely known.
And of course, either “Sexium” or “Hexium” would have had the Christian fundies screaming that the new chips were either designed for downloading porn (the Sexium) or to be a tool for witchcraft (the Hexium)
(I’m not joking. You should read some of the stuff I’ve come across where some fundies have called computers “satanic” after finding out that they have little programs called “daemons” in them.)
That work was done in Israel away from all the BS engineers had to put up with when working on the flag chip processor. When it was demonstrated to have better performance it became the flag ship processor line.
It wasn’t exactly anyone in marketing figuring this out – it was a decree by U.S. District Judge William Ingram saying that AMD had the right to sell a compatible chip and include “486” in the name that brought this fact home to Intel.
