Got a tip for us?
Let us know
Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.
Developer Delves Into Reasons Why Apple's M1 Chip is So Fast
- Thread starter MacRumors
- Start date
- Sort by reaction score
Unified memory. Specialized processing cores. That's two. Do you or do you not agree that moving the memory controller from the motherboard to the processor increased performance?
Those two cannot both be true. Changing the name of something does not change that something. It is the same thing under another name.
It’s a fair discussion, but the i7 model is $1,500. The M1 performing is performing pretty consistently at the $999 base model. Also, the software to compare isn’t really optimized for the M1 yet. I still think Apple has a leg up with the M1 considering it’s their first chip while Intel has been doing this for years.
Also if you look at the US market (which is where Apple tends to be somewhat more focused) the Mac market share is more like 25% than 8%. Considering that they have no low end machines (even the MBA price and fit and finish wise is not low end) they are a major player in their market.
Whaaa??? You mean there were other factors? I wish I had known: That there are other factors at play other than RISC versus CISC.
If only I had known we could have avoided all of this. Oh, wait.
So, IBM was only interested in high-performance variants that consumed more and more power? Really? You mean they wouldn't be interested in developing higher-performance processors that used equal or less power? They were solely focused on higher-performance at more and more power?
Motorola lost interest in capturing a market with higher-performing processor designs? Seems like a very poor business decision to ignore such a huge market.
This exactly. ^^^^^^
I worked for IBM and can tell you the interest was in LARGE as in HUGE systems. I was a part of the Power team and can tell you it was frustrating (as an Apple / Sony fan) we wanted to have scaled super chips on our desks and consoles. It was simply not the desired market for IBM.
I don't think they lost interest, they were leaking money like a sieve. BTW I also know a designer at Motorola (then moved to Freescale) that worked on PPC along with us. They wanted to keep it going as it was an oppertunity for them with Apple.
Why were they leaking money like a sieve? After all they were producing a RISC processor. I've read so much about how producing RISC processors is the way to go.
Typical response for this site. I didn't expect anything more.
Had nothing to do with that. It was largely related to how the cell phone market was going. Plus the culture and mismanagment played a role. I didn't work there, getting that last from my friend.
Maybe I could have: Unified memory. Specialized processing cores. That's two.
This is a prime example of the problem in these forums. People unwilling to listen.
RISC instruction set on ARM has been growing for years. However, the instructions are still fixed in length and they are not run in parallel as in CISC. Fundamentally - it is still RISC.
There are many factors that contribute to the success of M1 vs, say, Intel. RISC vs. CISC is a huge one. And RISC enables some of the other factors - it provides more degrees of freedom for the designers to play with in their physical design, for example. And it means they can dedicate transistors that would otherwise have to be spent on CISC weirdness for other things (like neural engines, etc).
Was this done through virtualisation? Is there a link with comparable metrics to other platforms?
I know that some version of the Amiga still exists. I also know Apple had little or nothing to do with its Commodore version failing. I owned an Amiga 1000, 2000, and 3000 between approximately 1985 and 1993. What killed it was that everyone wanted IBM compatibility (so if someone says Windows Compatibility they obviously weren’t there) and the fact that it was labeled a “game machine “. Today that’s a compliment, back then it was an insult. It didn’t help that Commodore really had no idea what they had or how to market it and what they needed software wise to compete. You could do video editing by the late 1980’s, the computer had built software for mixing pictures and video and sound. it was multi tasking years before other home computers. But it wasn’t IBM Compatible. I don’t know if it ever got out of prototype but there was an add-on called “Side-Car” that was supposed to give you the ability to run MS Dos programs. I don’t remember hearing if it would run Windows 1.0 to 3.11. By that time it’s sales were dwindling and when Windows 95 came out with all sorts of features that the Amiga had 10 years earlier every tech magazine was proclaiming what geniuses MicroSoft were.
By the time this was happening Steve Jobs had long left Apple, and IBM, Motorola and Apple were finishing the Power PC chip, where Apple now ran into the “it’s not compatible“ label but now at least it really was called ”Windows Compatible “ because you could buy motherboards and processors and memory from Computer Shopper magazine and build your own and weren’t locked into a big name brand. So the OS was more important than the commodity company that assembled the components and slapped a brand name on it.
I don’t know cmaier from jack but it’s obvious that he has designed chips and worked with both x86 and RISC designs, something other posters here on MacRumors have verified. What are your credentials?
If I recall correctly -- it was the manufacturing yield from IBM that wasn't keeping up not necessarily the technical attributes of RISC that forced the change. Now it seems the roles have reversed in that regard.
Plus -- Apple's unique implementation of ARM with their SOC design and the integration of software also makes the M1 competitive. Its not just a RISC vs CISC discussion
