Intel used to name each major processor release 286, the 386, the 486, etc. Then the next version was supposed to be called the 586, but their marketing department decided to rename it “pentium”, but to this day, if you pay attention when your computer starts up, you still see references to “X86” in the debugging output.
So how many major versions has Intel released since that 586/pentium? If they had kept the same naming convention, what number would we be on now? Like the 1286 or something?
You have to choose what separates each generation and as the processors got more varied and complex that dividing line got more blurred. Here is the historical list of the microarchitectures.
They didn’t start with -86. We’ve had the 4004, 8080, 8088 as well off the top of my head. I think we would have gone through -32 and -64 suffixes also, and then additional coding for multi-cores.
Yes but the poster asked specifically about starting at 286 and continuuing the count. The link that astro posted includes the major generation changes and includes the transition to 64 bit and multi cores. 2 x core and 4 x core versions of the same chipset are one generation.
It’s pretty arbitrary, Nehalem as far as I can tell was only a die shrink and minor changes from the original Core architecture, while the other generational changes are more clear. Anyway, Current Haswell processors are either 1086 or 1186 depending on how you count it.
So by my count, if the P5 was 5th generation, Ivy Bridge is 12th.
Note that that’s just for Pentiums, and there have been many other forks along the way, so the family tree for all Intel processors looks like the European royal lines.
Interesting side info: the reason Intel changed from 486 to Pentium is that the US Government ruled that numbers (like “486”) couldn’t be copyrighted. So they were unable to prevent other CPU maker from also calling their chips a 486 or 486-compatible. So Intel went to a made-up name like Pentium.
… and then kept the Pentium name because it had become a successful brand and, possibly, because the only reasonable “next steps” would be either Hexium or Sexium, neither of which would have been acceptable to marketing.
Probably mostly because the Pentium brand had become so successful.
So all of these “generations” are marketing first and foremost: There has never been a single line of development for x86 after the very early days (even the 80186 was off to one side, having been primarily used as an embedded processor) and, to programmers, it’s far more interesting to group chips by which features (ISA extensions, cache sizes, etc.) they have, regardless of the brand name stuck on them.
The grandfather of pretty much all PC chips is the 16-bit 8086, a major evolution over the foundational 8-bit 8080. The 8088 was the 8086 repackaged with an 8-bit external data path, which let IBM build a slower, cheaper machine for the first PC.
Other than regarding the 8080 as the Neandertal of the Intel family, I think you can safely ignore the other lines and earlier generations for this discussion.