Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Macintel: The End Is Nigh

mondaynote.com

Post reply on HN
Print view View original

71–80 of 156 posts

Re: Macintel: The End Is Nigh

#71

Post by ghaff »

fiatmoney wrote: post #29

"The aging x86 architecture is beset by layers of architectural silt accreted from a succession of additions to the instruction set... Because of this excess baggage, an x86 chip needs more transistors than its ARM-based equivalent" I wish people would stop saying this. On the inside, x86 CPUs are basically RISC. There is a translation layer from the publicly facing instruction layer to the internal representation. T…

We could also recycle lots of Intel's own verbiage from its own Itanium introduction :-) That didn't work out so well.

There are some interesting questions related to whatever convergence of tablets and laptops ends up happening. But it's not clear to me that especially favors ARM. And, it's also a hybrid sort of device that seems so obvious yet Apple finally got tablets to take off by taking a different path. So I'm skeptical about time frames at a minimum.

Re: Macintel: The End Is Nigh

#72

Post by rodgerd »

Symmetry wrote: post #45
ggreer wrote: post #26

What would happen to the cost, battery life, and size of an A10-powered MacBook Air? It would be worse in almost every aspect. The cost wouldn't change much, and Apple wouldn't profit much more from the switch. The cited CPU cost is the suggested retail price . Apple's volume lets them negotiate huge discounts. Battery life would go up a bit, but battery life is already pretty good on Apple laptops. My 2013 11" Air g…

You say that ARM chips are 10x slower than Haswells, but the difference isn't actually that large anymore and for Apple's A7 it's around 3x versus a normal desktop chip like the 4770 and only 1.7X versus the 4250U you'd find in a Macbook Air[1]. Diminishing returns is a constant factor in engineering and I'd expect it to be very hard for Apple to make up the remaining distance between themselves and Intel, but we sho…

> for Apple's A7 it's around 3x versus a normal desktop chip like the 4770 and only 1.7X versus the 4250U you'd find in a Macbook Air

That would still be noticeably slower for native software, and unbearable for anything run under emulation. It would also be very painful for use cases such as people wanting occasional Windows applications or Linux VMs.

Re: Macintel: The End Is Nigh

#74

Post by rodgerd »

ggreer wrote: post #47

While I think the article's prediction is incorrect (see my other comment for why), it did give me an idea: Why doesn't Apple design their own x86 CPU? Their work on the Ax series has probably taught them quite a bit that could be ported over to x86-land. Also, Apple could leave out x86 cruft they don't use: legacy addressing modes, PAE, etc. And of course, they could design the CPU specifically for their products in…

> Why doesn't Apple design their own x86 CPU?

There's the small matter of getting an x86 license. Which would mean buying Intel or AMD, since it's vanishingly unlikely that Intel will be handing them out any time soon.

Re: Macintel: The End Is Nigh

#75

Post by Symmetry »

higherpurpose wrote: post #61
Symmetry wrote: post #45

Earlier quoted context omitted.

You say that ARM chips are 10x slower than Haswells, but the difference isn't actually that large anymore and for Apple's A7 it's around 3x versus a normal desktop chip like the 4770 and only 1.7X versus the 4250U you'd find in a Macbook Air[1]. Diminishing returns is a constant factor in engineering and I'd expect it to be very hard for Apple to make up the remaining distance between themselves and Intel, but we sho…

The difference is actually smaller, since we're comparing to a 1.3 Ghz Air, not a 3.3 Ghz (or whatever) desktop chip.

The 4250U I mentioned is the one in the 1.3 Ghz Air.

Re: Macintel: The End Is Nigh

#76

Post by Symmetry »

rodgerd wrote: post #72
Symmetry wrote: post #45

Earlier quoted context omitted.

You say that ARM chips are 10x slower than Haswells, but the difference isn't actually that large anymore and for Apple's A7 it's around 3x versus a normal desktop chip like the 4770 and only 1.7X versus the 4250U you'd find in a Macbook Air[1]. Diminishing returns is a constant factor in engineering and I'd expect it to be very hard for Apple to make up the remaining distance between themselves and Intel, but we sho…

> for Apple's A7 it's around 3x versus a normal desktop chip like the 4770 and only 1.7X versus the 4250U you'd find in a Macbook Air That would still be noticeably slower for native software, and unbearable for anything run under emulation. It would also be very painful for use cases such as people wanting occasional Windows applications or Linux VMs.

Right, using an A7 in a Macbook would be totally crazy. The idea is that Apple might create an A8 that's about as fast as the current Macbook Air and use that. Getting as much performance increase going from A7 -> A8 as they did going A6 -> A7 strikes me as implausible, and it would probably be a bad idea to switch architectures anyways. But it's not totally crazy.

Re: Macintel: The End Is Nigh

#77

Post by ANTSANTS »

Symmetry wrote: post #52
fiatmoney wrote: post #29

"The aging x86 architecture is beset by layers of architectural silt accreted from a succession of additions to the instruction set... Because of this excess baggage, an x86 chip needs more transistors than its ARM-based equivalent" I wish people would stop saying this. On the inside, x86 CPUs are basically RISC. There is a translation layer from the publicly facing instruction layer to the internal representation. T…

It actually is true, though it's easy to exaggerate how important it is. Decoding an x86 instruction into the RISCish internal format isn't hard, but decoding 4 at once can be quite a challenge. There's no way to look at a single byte of the incoming instruction stream and figure out if it's the start of an instruction or not, you need to do a fair amount of decoding on each byte of the chunk that you fetch from I$ b…

OTOH you can cache the decoded microops (IIRC modern Intel chips do this, while modern AMD and the original Pentium cached instruction boundaries), and in thumb mode ARM is also variable length.

Re: Macintel: The End Is Nigh

#78

Post by ggreer »

Symmetry wrote: post #45
ggreer wrote: post #26

What would happen to the cost, battery life, and size of an A10-powered MacBook Air? It would be worse in almost every aspect. The cost wouldn't change much, and Apple wouldn't profit much more from the switch. The cited CPU cost is the suggested retail price . Apple's volume lets them negotiate huge discounts. Battery life would go up a bit, but battery life is already pretty good on Apple laptops. My 2013 11" Air g…

You say that ARM chips are 10x slower than Haswells, but the difference isn't actually that large anymore and for Apple's A7 it's around 3x versus a normal desktop chip like the 4770 and only 1.7X versus the 4250U you'd find in a Macbook Air[1]. Diminishing returns is a constant factor in engineering and I'd expect it to be very hard for Apple to make up the remaining distance between themselves and Intel, but we sho…

I think GeekBench is a poor benchmark, since it's measuring peak CPU performance in a narrow domain, but the table you cited shows that the 4770 is 6.5x faster than the A7 (16774 vs 2564). Even at the same frequency and core count, Haswell is 1.8x faster in the benchmark.

Whatever the exact number, it doesn't invalidate the point I was making: Switching to ARM will ruin performance for x86 programs, which need to be emulated. ARM has to be significantly faster than x86, or else the emulation overhead will make users hate the transition.

This is a huge challenge for the Mac Pro and rMBP, which are all about performance. Generally, a given microarchitecture can only scale TDP by a factor of 10 by changing frequency and number of cores. Satisfying the pro line would require a new microarchitecture. Considering an x86 to ARM transition would take 1-2 years, that's asking a lot of Apple's chip designers. If Apple was going to go through all the trouble of building a new ARM core for high-performance, high-TDP applications, they might as well build their own x86 CPU. It would save a transition and the necessary emulation overhead. They could roll it out for the MacBook Airs first, then address the Pro line at their leisure (or not and just use Intel CPUs).

Re: Macintel: The End Is Nigh

#79

Post by PhantomGremlin »

jblow wrote: post #51
fiatmoney wrote: post #29

"The aging x86 architecture is beset by layers of architectural silt accreted from a succession of additions to the instruction set... Because of this excess baggage, an x86 chip needs more transistors than its ARM-based equivalent" I wish people would stop saying this. On the inside, x86 CPUs are basically RISC. There is a translation layer from the publicly facing instruction layer to the internal representation. T…

Yeah, as soon as the guy started saying this I discounted the whole article. Yes, maybe Apple will switch to A7 chips, but this author sure doesn't know enough about processors to have any kind of a privileged viewpoint. If one doesn't even know what uops are, and doesn't have a mental estimate of what percentage of i7 silicon is devoted to the instruction decoder, one doesn't get to write articles comparing Intel to…

> I note the author is Jean-Louis Gassee

Yeah, not sure why anyone is interested in what Gassee has to say. He used to be "head of Apple's advanced product development", but was forced out for failing to deliver.[1]

Also, he should get the small details right. He claims in the article

   The x86 nickname used to designate Wintel chips
   originates from the 8086 processor introduced
   in 1978 – itself a backward-compatible extension
   of the 8088…)
This is exactly backwards. The 8086 was first, ahead of the 8088. Not sure why Gassee even felt he needed to throw that detail into the article (unless he's getting paid by the word), but if he's going to say it, he should at least check Wikipedia first.[2]

[1] http://en.wikipedia.org/wiki/Jean-Louis_Gass%C3%A9e [2] http://en.wikipedia.org/wiki/Intel_8086

Re: Macintel: The End Is Nigh

#80

Post by wmf »

higherpurpose wrote: post #64
fiatmoney wrote: post #29

"The aging x86 architecture is beset by layers of architectural silt accreted from a succession of additions to the instruction set... Because of this excess baggage, an x86 chip needs more transistors than its ARM-based equivalent" I wish people would stop saying this. On the inside, x86 CPUs are basically RISC. There is a translation layer from the publicly facing instruction layer to the internal representation. T…

Intel chips even support 16-bit processing. How is that not baggage?

It is baggage, but it costs very little.
Post reply on HN