Apple Mac Going to ARM RISC
-
@Dashrender said in Apple Mac Going to ARM RISC:
I know it wasn't on PC
Remember, Apple Mac has been the most "PC" of any computer since 2003.
Now, and this drives me absolutely crazy, for the first time in 17 years, the ARM Macs will not be PC again, so all the people incorrectly using PC to refer to something that isn't necessarily PC when Mac is the most PC thing ever made since the original IBM PCs, will suddenly sound correct again.
-
PC refers to machines running IA32 and now AMD64 architecture with the PC infrastructure. So the vast majority of Linux and Windows, and all Mac until now.
So phones have not been PCs. Raspberry Pi is not PC. Anything with Power, ARM, Sparc, RISC-V and similar chips are not PC. When Windows runs on AMD64 it is PC, when it runs on ARM it is not.
-
@scottalanmiller said in Apple Mac Going to ARM RISC:
PC refers to machines running IA32 and now AMD64 architecture with the PC infrastructure. So the vast majority of Linux and Windows, and all Mac until now.
So phones have not been PCs. Raspberry Pi is not PC. Anything with Power, ARM, Sparc, RISC-V and similar chips are not PC. When Windows runs on AMD64 it is PC, when it runs on ARM it is not.
So, it's the CPU architecture that determines whether or not a computer is personal?
-
@Obsolesce said in Apple Mac Going to ARM RISC:
So, it's the CPU architecture that determines whether or not a computer is personal?
No, it's what determines if it is a PC. IBM PC is an IBM/Intel designed computer architecture that was originally used with CP/M and later other operating systems were written for it. It's the most well known computing architecture standard. While PC standard for Personal Computer, don't confuse it with the more general personal computer. Any computer that is personal is a personal computer, but nothing in being a personal computer implies that it is a PC and nothing in PC actually means it's personal. Every typical server you use is a PC, but the average personal computer is not.
A server is like 95% certain to be a PC today (AMD64 PC architecture) but a personal computer is only like 30% likely to be a PC (because most personal computers are ARM based mobile devices.)
Chromebooks famously come in both PC and non-PC variants.
-
@scottalanmiller said in Apple Mac Going to ARM RISC:
So, it's the CPU architecture that determines
It's more than that. You can make an IA16, IA32, or AMD64 CPU based system that is not a PC, but you'd be engineering all your own parts and connections. It's the PC architecture that makes it possible to have interchangeable parts, the PCI bus and stuff like that. So some embedded systems that use those processors opt to not build a PC, but it's rare because once you diverge from PC you generally do so enough that you are stuck writing your own operating system and that's really expensive.
The PC standard is what makes those processors cheap and standard. Pretty much you choose PC first, then you choose which PC compatible processor you want, not the other way around as it would rarely, if ever, make sense. If you dont want PC, you don't likely want an expensive PC focused processor.
-
@scottalanmiller said in Apple Mac Going to ARM RISC:
@Dashrender said in Apple Mac Going to ARM RISC:
I know it wasn't on PC
Remember, Apple Mac has been the most "PC" of any computer since 2003.
Now, and this drives me absolutely crazy, for the first time in 17 years, the ARM Macs will not be PC again, so all the people incorrectly using PC to refer to something that isn't necessarily PC when Mac is the most PC thing ever made since the original IBM PCs, will suddenly sound correct again.
/sigh - I know.
-
In the old days, there used to be an actual PC. PC is the name of a specific computer model from IBM. Computers that used that architecture, but weren't the PC itself, were called PC-compatible. The IBM PC, aka the 5250, was the first of the architecture that we use today. It was designed around CP/M, but DOS was quickly made for it (DOS is a CP/M clone anyway.)
Once IBM discontinued making their own PC line, people stopped using the term PC-compatible and everything that was "compatible" became simply known as a PC.
The term PC in the name of IBM's PC was to stand for "personal computer", but they weren't personal computers, they were computers generally used by individual in a company. They were very business oriented, not personal like everyone things of today. They were in competition with those primarily. The reference today is to the product, not the incorrect reference that that product made.
Even back in 1984 there were non-PC computers that ran Intel processors and would run CP/M or DOS or both. So even from the very beginning PC was very specifically one thing while the OS and even chips were not part of it. Being PC or PC-compatible never referred to the OS in any way, only to the hardware. An OS would have to be written for PC/PC-compatible to run on it, but any OS that runs on it is equally a PC OS. And Windows was the third OS to be famous on PC, after DOS, which came after CP/M. But many others, even from Microsoft like Xenix, were PC OSes meant only for that architecture.
And Windows, almost always, and DOS absolutely always, has been available in both PC and non-PC variants.
-
@scottalanmiller said in Apple Mac Going to ARM RISC:
@scottalanmiller said in Apple Mac Going to ARM RISC:
So, it's the CPU architecture that determines
It's more than that. You can make an IA16, IA32, or AMD64 CPU based system that is not a PC, but you'd be engineering all your own parts and connections. It's the PC architecture that makes it possible to have interchangeable parts, the PCI bus and stuff like that. So some embedded systems that use those processors opt to not build a PC, but it's rare because once you diverge from PC you generally do so enough that you are stuck writing your own operating system and that's really expensive.
The PC standard is what makes those processors cheap and standard. Pretty much you choose PC first, then you choose which PC compatible processor you want, not the other way around as it would rarely, if ever, make sense. If you dont want PC, you don't likely want an expensive PC focused processor.
I like NIST's definition better.
I think I'll stick to that until you find some other better documentation. I can't really find anything official that goes into the kind of background you do.
-
To add to the confusion, Apple decided to license Power from IBM and make their own Power processors (actually via the Motorola foundry) that they branded PowerPC, in that case the PC stood for personal computer again, but had no association with PC architecture and was a completely competing product.
But PC was an architecture, and PowerPC was a specific brand of Power architecture. So even when you had PowerPC, the architecture was Power, not PowerPC. PowerPC was more like an AMD Sempron - a low performance, cheap, third party copy of someone else's original.
But PC was all the rage as something to say in the late 1990s, so Apple attempted to capitalize on its popularity by adding it to a brand name. But Power was a bad design for personal computing / desktop needs and PowerPC was an anemic offering in the Power lineup and it all failed. By 2003 Apple decided to finally ditch having their own hardware vertical and embraced the PC architecture 100%. So starting in 2003, the Apple Mac lineup became a PC in every way, and because they were the only brand that absolutely required Intel-made processors, and the original PC was Intel, one could argue that they were the most completely PC computer made since the original IBM PC line.
-
@Obsolesce said in Apple Mac Going to ARM RISC:
@scottalanmiller said in Apple Mac Going to ARM RISC:
@scottalanmiller said in Apple Mac Going to ARM RISC:
So, it's the CPU architecture that determines
It's more than that. You can make an IA16, IA32, or AMD64 CPU based system that is not a PC, but you'd be engineering all your own parts and connections. It's the PC architecture that makes it possible to have interchangeable parts, the PCI bus and stuff like that. So some embedded systems that use those processors opt to not build a PC, but it's rare because once you diverge from PC you generally do so enough that you are stuck writing your own operating system and that's really expensive.
The PC standard is what makes those processors cheap and standard. Pretty much you choose PC first, then you choose which PC compatible processor you want, not the other way around as it would rarely, if ever, make sense. If you dont want PC, you don't likely want an expensive PC focused processor.
I like NIST's definition better.
I think I'll stick to that until you find some other better documentation. I can't really find anything official that goes into the kind of background you do.
Yeah, but NIST just making up a new definition for an industry term means nothing. That they are willing to lie doesn't make it true.
-
@scottalanmiller said in Apple Mac Going to ARM RISC:
@Obsolesce said in Apple Mac Going to ARM RISC:
@scottalanmiller said in Apple Mac Going to ARM RISC:
@scottalanmiller said in Apple Mac Going to ARM RISC:
So, it's the CPU architecture that determines
It's more than that. You can make an IA16, IA32, or AMD64 CPU based system that is not a PC, but you'd be engineering all your own parts and connections. It's the PC architecture that makes it possible to have interchangeable parts, the PCI bus and stuff like that. So some embedded systems that use those processors opt to not build a PC, but it's rare because once you diverge from PC you generally do so enough that you are stuck writing your own operating system and that's really expensive.
The PC standard is what makes those processors cheap and standard. Pretty much you choose PC first, then you choose which PC compatible processor you want, not the other way around as it would rarely, if ever, make sense. If you dont want PC, you don't likely want an expensive PC focused processor.
I like NIST's definition better.
I think I'll stick to that until you find some other better documentation. I can't really find anything official that goes into the kind of background you do.
Yeah, but NIST just making up a new definition for an industry term means nothing. That they are willing to lie doesn't make it true.
I think I'd be better off sticking to a defined industry term by NIST than something one person says.
-
@Obsolesce their definition isn't even a definition... they call it something running a PC OS. WTF is a PC OS?
However, if you look at their definition and think about it, you can tell that they are TRYING to tell you what I just said. PC OS is pretty obviously intended to mean an OS written to run on the PC architecture, and their examples are just that. Vista, and Mac OSX run ONLY on PC architecture, and Linux was originally written for it (then branched out.)
So while this is an incredibly sloppy and dumbed down definition that leaves all the hard work in the definition of "What is a PC OS", you can tell from their wording and their examples, that it's exactly what I just said. I just provided the history and "why" so that it all makes sense.
And notice, like I had said, that Personal Computer is capitalized, PC = Personal Computer, but PC != personal computer.
-
@Obsolesce said in Apple Mac Going to ARM RISC:
@scottalanmiller said in Apple Mac Going to ARM RISC:
@Obsolesce said in Apple Mac Going to ARM RISC:
@scottalanmiller said in Apple Mac Going to ARM RISC:
@scottalanmiller said in Apple Mac Going to ARM RISC:
So, it's the CPU architecture that determines
It's more than that. You can make an IA16, IA32, or AMD64 CPU based system that is not a PC, but you'd be engineering all your own parts and connections. It's the PC architecture that makes it possible to have interchangeable parts, the PCI bus and stuff like that. So some embedded systems that use those processors opt to not build a PC, but it's rare because once you diverge from PC you generally do so enough that you are stuck writing your own operating system and that's really expensive.
The PC standard is what makes those processors cheap and standard. Pretty much you choose PC first, then you choose which PC compatible processor you want, not the other way around as it would rarely, if ever, make sense. If you dont want PC, you don't likely want an expensive PC focused processor.
I like NIST's definition better.
I think I'll stick to that until you find some other better documentation. I can't really find anything official that goes into the kind of background you do.
Yeah, but NIST just making up a new definition for an industry term means nothing. That they are willing to lie doesn't make it true.
I think I'd be better off sticking to a defined industry term by NIST than something one person says.
Well, that's a really bad logic. Because I provided exactly why it is what it is, and have had the same answer for decades. The NIST example is new and obviously isn't a useful or reliable source. It has neither logic, nor history, nor reliably applicability. It's clear from looking at it that it's not a viable description.
That said, if you use it in any meaningfully logical way, it leads you right back to what I original said. So you end up agreeing with me, except with skipping the understanding of why - which is the underlying problem in IT. Everyone wants to memorize definitions or simple answers instead of understanding what terms, technologies, or whatever actually mean. If you understand what PC was from the beginning, which is actually really simple, then knowing how it applies later is easy. If you try to memorize a quick answer that doesn't address the actual factors making it true, then it's easy to get confused or lost or need to constantly have to "figure out" when it applies or doesn't.
-
For example... the original Mac OSX in 2000 was not a PC OS. But in 2003, they released the PC OS version of it. Now that they are going to ARM, it'll be like iOS and won't be a PC OS any more.
The NIST example tracks what I said, it just doesn't usefully provide enough information to be used on its own. So you can say you'll "just use it", except you can't, because it's incomplete and required my explanation to make sense.
-
@scottalanmiller said in Apple Mac Going to ARM RISC:
@Obsolesce said in Apple Mac Going to ARM RISC:
@scottalanmiller said in Apple Mac Going to ARM RISC:
@Obsolesce said in Apple Mac Going to ARM RISC:
@scottalanmiller said in Apple Mac Going to ARM RISC:
@scottalanmiller said in Apple Mac Going to ARM RISC:
So, it's the CPU architecture that determines
It's more than that. You can make an IA16, IA32, or AMD64 CPU based system that is not a PC, but you'd be engineering all your own parts and connections. It's the PC architecture that makes it possible to have interchangeable parts, the PCI bus and stuff like that. So some embedded systems that use those processors opt to not build a PC, but it's rare because once you diverge from PC you generally do so enough that you are stuck writing your own operating system and that's really expensive.
The PC standard is what makes those processors cheap and standard. Pretty much you choose PC first, then you choose which PC compatible processor you want, not the other way around as it would rarely, if ever, make sense. If you dont want PC, you don't likely want an expensive PC focused processor.
I like NIST's definition better.
I think I'll stick to that until you find some other better documentation. I can't really find anything official that goes into the kind of background you do.
Yeah, but NIST just making up a new definition for an industry term means nothing. That they are willing to lie doesn't make it true.
I think I'd be better off sticking to a defined industry term by NIST than something one person says.
Well, that's a really bad logic. Because I provided exactly why it is what it is, and have had the same answer for decades. The NIST example is new and obviously isn't a useful or reliable source. It has neither logic, nor history, nor reliably applicability. It's clear from looking at it that it's not a viable description.
That said, if you use it in any meaningfully logical way, it leads you right back to what I original said. So you end up agreeing with me, except with skipping the understanding of why - which is the underlying problem in IT. Everyone wants to memorize definitions or simple answers instead of understanding what terms, technologies, or whatever actually mean. If you understand what PC was from the beginning, which is actually really simple, then knowing how it applies later is easy. If you try to memorize a quick answer that doesn't address the actual factors making it true, then it's easy to get confused or lost or need to constantly have to "figure out" when it applies or doesn't.
Yeah I see, it sets you up for a loop. (the NIST definition)
Where can I find the "PC Standard"?
-
Here is an example of a multi-architectural company...
https://www.fujitsu.com/global/about/corporate/history/products/computer/server/
Fujitsu, who makes or made both Sparc and IA architecture machines, points out that IA servers are PC servers, because they are PC architecture, as opposed to Sparc architecture servers that are not PC. However, both run the same OSes.
A key problem with the definitions from NIST is that long ago, operating systems were generally limited to a single architecture and so the two could be often associated to some degree. In the years since, that has changed and today nearly all operating systems run on many, if not most, architectures. Essentially every OS will run on PC due to how popular and standard it is. Raspberry Pi (aka RP32 and RP64) is decently common now. Apple's iPhone / iPad architecture is huge as well.
So once upon a time, definitions could be written with those assumptions to make it easy for people to identify a PC without understanding the tech. But today, you can't. And that oversimplication clearly makes it make no sense.
Android can run on ARM, PC, and others. Linux runs on everything. Solaris will run on PC. Windows will run on Itanian (IA64). Mac OSX will run on ARM. and so forth. Things assumed to not be PC OSes now are (optionally) and nearly every thing assumed to be a PC OS will (optionally) run on something other than a PC.
But something is only a PC OS when it runs on a PC.
-
@Obsolesce said in Apple Mac Going to ARM RISC:
Where can I find the "PC Standard"?
Well the original was defined by IBM in ~1981 when they invented it. It wasn't originally meant to be a standard, but became one. So it's a de facto standard that stems from being compatible (or backwards compatible) with the original PC and its PC offspring.
-
I get all of what Scott is saying - regardless to Scott's research - lay people us PC (including almost all journalist) to mean a computer running Windows - they might even mean AMD64 based Windows instead of ARM based Windows - but even that's a stretch.
Let me say again - Scott's history lesson - again - is correct/likely correct, but that has no baring on how the general public use the term.
PC = windows
Mac = mac
period.
no lay person will ever call ca Mac a PC. -
So PC works from a reference standard. But IBM passed the reigns, somewhere around the late 1980s or 1990 to Intel. Intel defines the reference standard today as part of their IA32/AMD64 reference designs. AMD just follows suit. AMD defines the processors, but Intel defines the interconnects that make the larger PC ecosystem. To some degree, lots of vendors work together, of course, including Apple in no small part.
So things like USB, IEEE/1394, SATA, PCIe and so forth are now part of what is considered the PC standard. Other architectures can and do use them, but they generally originate from PC and get co-opted. But now they work with the processor is part of the PC standard.
It's not something you normally see outside of electrical engineering circles because it's at the motherboard and chip manufacturing level and broadly irrelevant to end users, even when those end users are senior IT people. We just know it's compatible with whatever OS we want. And no one sells non-PC Intel / AMD boards these days, so that makes it simple.
-
@Dashrender said in Apple Mac Going to ARM RISC:
no lay person will ever call ca Mac a PC.
Lots do and always have. In fact that's always been a huge part of the problem because Macs were always personal computers, whereas PCs rarely were for the first eight years or so. So Macs were always heavily associated with the term by confused people long before they became a PC.
Then when Apple moved from M68K Mac architecture (which was different than Atari and Commodore's M68K architectures) to PowerPC, they did so to some degree intentionally to get Mac people to refer to them as PCs for marketing reasons, and it worked like crazy. Macs used to always be called PCs, incorrectly. I used to have this same discussion, in reverse, constantly.
When Macs actually moved to PC it's true, the absolutely confused Mac user crowd actually starting calling them PCs less than every before. So my guess is that when they leave PC architecture to go to ARM that people will call them PC more then than now.
But lay people call them PCs all of the time, and NIST's definition 100% supports that. The idea that Windows = PC and nothing else does is something I hear, but not all that often. PC as a general catch all for all personal computers, that I hear a lot and includes Mac every time. The problem there is that it also starts to include iPads and such that is wrong.
That Surfaces are PCs and iPads are not adds to the confusion.
