Register a SA Forums Account here!
JOINING THE SA FORUMS WILL REMOVE THIS BIG AD, THE ANNOYING UNDERLINED ADS, AND STUPID INTERSTITIAL ADS!!!

You can: log in, read the tech support FAQ, or request your lost password. This dumb message (and those ads) will appear on every screen until you register! Get rid of this crap by registering your own SA Forums Account and joining roughly 150,000 Goons, for the one-time price of $9.95! We charge money because it costs us money per month for bills, and since we don't believe in showing ads to our users, we try to make the money back through forum registrations.
 
  • Post
  • Reply
P.N.T.M.
Jan 14, 2006

tiny dinosaurs


Fun Shoe

It depends on what Apple cares about. I'd like to think Apple cares more about its image than anything else, but they surely don't. What Apple seems to really care about is: "There shall be no waste."

Primarily,

No waste of space.

No waste of power.


They've cut their product lines down to 6 categories. The Mac Pro is a re-envisioned work-station for the 21st century. The MBP is an expensive solution unto itself. The MBA is Apple's idea of what a normal person's laptop should be, and the iMac is their desktop solution. The iPhone and iPad are obviously the phone and tablet offerings.


What could ARM bring to Apple? More market share when prices get a sliver taken off, increased control through in-house specs for all processors, and better offerings in the future as they further develop their own technology. Right now, they have to reserve space and energy to power Intel's hardware, which must have recently become too much overhead for Apple to accept.

Adbot
ADBOT LOVES YOU

bull3964
Nov 18, 2000

DO YOU HEAR THAT? THAT'S THE SOUND OF ME PATTING MYSELF ON THE BACK.




That's the thing though, the latest Intel low power chips have proven themselves to be MORE power efficient than many ARM chips, so I really don't buy the power thing. Apple isn't going to out-engineer Intel at this point.

You may view it as 6 market segments, but really it's 2. iOS and OSX. Thus far, there has been a clear delineation between the two. The former is a touch optimized interface with no legacy app support and the later eschews touch entirely and relies on trackpads and keyboards.

An ARM notebook would be somewhere in-between. They would either have to branch OSX for ARM and then have a version of their OS that's completely incapable of running a huge swath of software and increased market confusion for nebulous efficiency gains or they would put iOS on it. However, that would be adapting iOS for a hardware target completely unlike anything that has used it before. Touch would be de-emphasized due to the form factor which has serious implications for the usability of iOS.

It's adding a serious level of complexity to their market line for goals that no one is asking for right now.

evilweasel
Aug 24, 2002



P.N.T.M. posted:

What could ARM bring to Apple? More market share when prices get a sliver taken off, increased control through in-house specs for all processors, and better offerings in the future as they further develop their own technology. Right now, they have to reserve space and energy to power Intel's hardware, which must have recently become too much overhead for Apple to accept.

The issue is, Intel is starting to massively spend on doing better on power consumption. And the thing people forget about ARM is it doesn't hold a candle to the actual horsepower of an Intel Core chip: ARM is doing so well because there's so many situations where that horsepower is just not needed and you'd rather have less power and less power use. However, macbooks are premium laptops and I'm not sure they'd want to make that tradeoff, especially with Intel rapidly eliminating ARM's power advantage.

Hendrik
Feb 5, 2009


Reduced power requirements boots battery life. If the power efficiency of ARM can be maintained with slightly enhanced performance we could have another wave of netbook equivalents. An expensive device with an impressive battery life could fit into Apple's product line.

necrobobsledder
Mar 21, 2005
Lay down your soul to the gods rock 'n roll

Nap Ghost

If battery capacities get some huge improvements in upcoming years, a lot of the ARM v. x86 debate will be of little concern anyway... but the price floor of the chips will become more important, and this is where ARM will win by default compared to Intel in the same way Linux didn't affect Windows at all - it killed off commercial Unixes. Intel doesn't want to be relegated to the datacenter / server room because they don't want to be the mainframe CPU manufacturer.

I'm more concerned about progress in software than hardware though, especially on the bloated enterprise side. If we keep making horribly bullshit expensive software like the healthcare.gov site, that'll really discourage the primary drivers that fuel the continued progress on these improvements. Every other enterprise trying to build their own crappy version of EC2 and failing miserably from stepping on their own toes does not bode well for many companies that keep buying lots of Intel hardware.

The sheer amount of legacy software that people will want to keep running even in business cases will hardly matter at a point to the economy because their consideration is about as important as the people that keep using carbon copy paper for credit card transactions (read: they're not a source of revenue growth, a tiny fraction of revenue, and extremely stable / stubborn - why try to change them?). Our equities market doesn't really give a rip about stable business, they're only concerned with growth and will shift money accordingly.

mobby_6kl
Aug 9, 2009

"You are the best poster... do not let anyone say otherwise."


The Haswell gains are very impressive on the mobile and Broadwell should be as well, from what I've seen. More importantly, IMO, are the diminishing returns - Intel can keep cutting the power consumption by 30% each generation, but if the displays keep consuming what they do, the marginal improvement from going ARM are going to be, well, marginal at best. Then there's the whole Bay Trail thing, so I really have no idea what all this is supposed to be about.

Factory Factory
Mar 19, 2010

This is what
Arcane Velocity was like.


Hendrik posted:

Reduced power requirements boots battery life. If the power efficiency of ARM can be maintained with slightly enhanced performance we could have another wave of netbook equivalents. An expensive device with an impressive battery life could fit into Apple's product line.

Most of Apple's devices are already expensive devices with impressive battery life.

We're already getting a new wave of netbooks with the new Atoms, as well as with ARM and x86 Chromebooks.

ARM really doesn't have an inherent advantage over x86 any more. It used to be that x86 was built for speed and ARM was built for power and neither paid terribly much attention to the other. Some folks said that you could never make a low-power x86 core, first because it was CISC and second, once it became RISC + x86 decoder, because the decoder was too complex. These folks were wrong, and the die area difference between an x86 decoder and an ARM decoder has become almost trivial. Over the past decade or so, x86 has gotten lower-power, and ARM has gotten higher performance. And they kinda met in the middle.

ARM has ramped up performance immensely, and the high-end SoCs struggle to remain within their thermal envelopes. Current high-end ARM SoCs use fine-grained power gating and turbo implementations that, pre-Haswell, would have made Intel blush. It's pretty much at the practical limit of performance within phone and tablet power envelopes for the 28nm process node. If you include the 64-bit ARM uarch that debuted with Apple's A7 SoC, practically all the fancy features have been added on - out-of-order execution, internal parallelism, hardware prefetch, branch prediction, more and more advanced SIMD FPUs,...

You stick one of the high-end SoCs in a phone, and boy howdy does it get hot and burn through battery fast.

The new Atoms bring better CPU performance in the same power envelope, aided by Intel's 22nm process. The GPU side is a bit weak because Intel's graphics are a bit weak, but a new GPU uarch is up for next year. When you compare a Bay Trail SoC to, say, an SoC based on Cortex A15 or Krait 300, you get some tellingly similar design choices - low-parallelism CPU cores (dual-issue), similar clock speeds and core counts, out-of-order execution, fine-grained power gating, wide turbo clocking (50% or more difference between base and max turbo, with something like a 2x difference between short-term TDP and long-term SDP), security and crypto, GPU features (albeit with varying horsepower), blah blah blah. If you're familiar with AMD parts, that's Kabini, too (e.g. the A4-5000, what's used in the Xboner and PS4).

It kinda seems like there's a Right Way to do a power/performance balanced CPU, and everyone's found it.

Gwaihir
Dec 8, 2009



Hair Elf

Hendrik posted:

Reduced power requirements boots battery life. If the power efficiency of ARM can be maintained with slightly enhanced performance we could have another wave of netbook equivalents. An expensive device with an impressive battery life could fit into Apple's product line.

The X86 power disadvantage has been a myth debunked a pretty decent number of times at this point, I think. The generally more powerful x86 cores end up using less total power to perform given tasks due to the race to sleep paradigm that's been popularized in the time since CPUs have been able to take advantage of Turbo modes and power gating on idle cores. The current arm platform power advantages come from things other than the CPU cores themselves, whether it's RAM, storage, cellular basebands, or other chipset items. That's incidentally why Intel is focusing so much on reducing total platform power across all devices, as well as continuing to move more pieces on to the CPUs themselves.

I really wish Anand would re-do his detailed power benchmarking article from the beginning of this year with the latest Intel and Qualcomm (Or Apple, since he loving loves Apple everything) chips. It's one of the best comprehensive looks at power usage and performance efficiency out there, but there's been some huge leaps in the stuff available since then- He was looking at the old awful Atom arch, Krait, and Tegra 3 chips.

efb by Factory Factory.

necrobobsledder
Mar 21, 2005
Lay down your soul to the gods rock 'n roll

Nap Ghost

mobby_6kl posted:

if the displays keep consuming what they do, the marginal improvement from going ARM are going to be, well, marginal at best.
IGZO screens are in production test run phases now and they use a fraction of power of what current displays use. It may be a sort of stop-gap until OLED screens are more prominent / cost-effective, but given the hurdles that OLEDs are hitting, IGZO is probably Good Enough until we get the literally painted-on OLED screen surfaces that can turn almost anything into a display. But that point we may have bigger problems on our hands than "omg, my phone doesn't last more than a week without a charge."

http://www.idownloadblog.com/2013/0...013-sharp-igzo/

JawnV6
Jul 4, 2004

So hot ...

Hendrik posted:

Reduced power requirements boots battery life.
It's not that simple anymore. Nobody's making anything that runs ON or OFF. Everything's using dynamic frequency/voltage scaling so things like "race to idle" complicate the power/performance/battery picture.

Factory Factory posted:

It kinda seems like there's a Right Way to do a power/performance balanced CPU, and everyone's found it.
I see what you mean, but there's a lot of interesting work left to do in the area. Power management went from being a single FSM to it's own micro and has a lot more headroom to grow into.

I think it's kinda stupid, but Heterogenous compute cores are one possible power/perf solution that isn't guaranteed or discarded yet. It seems pretty stupid to ship 40% more silicon just to give a chip a few more DVFS points to scale to, but there are shipping designs using it.

Alereon
Feb 6, 2004

Dehumanize yourself and face to Trumpshed

College Slice

A few general points:

Why would Apple want to switch to ARM? Primarily so they can design and thus have total control over the SoCs in their own products. Intel has disappointed them on a number of occasions, for example Ivy Bridge supported something like Haswell's eDRAM and Apple was very excited about this (see how they bought all the R-series Haswells). Intel chose not to launch the eDRAM product and kind of left Apple in the lurch. Intel also still doesn't prioritize graphics (or drivers) in the way that Apple wants.

But isn't ARM too slow? Too slow for what? It would be a long time before Apple could switch Macbook Pros (or iMacs or Mac Pros) to ARM, but their low-end devices like the Macbook Air and the now-retired Macbooks (and maybe even Mac Mini) are targets. ARM processors CAN scale up in performance, that is exactly what Apple has done with the Cyclone core in the A7 CPUs. More importantly, Apple would redesign their product stack around the strengths of the CPU lineup they actually had available to them.

But isn't Intel too fast? Not <25W. Intel has made some impressive scaling gains with Haswell and will make more, but the 15W processors are being pushed out of their optimal range. We know Intel thinks four low-power Silvermont cores are better than two underclocked Haswell cores in many situations, and it looks like Cyclone can be competitive with Silvermont. >25W Intel's performance advantage is compelling but pricing reflects this. If Apple wanted to get into a lower-priced market they would want to push their bill of materials cost as low as possible without sacrificing the user experience.

But aren't high-end ARM SoCs inefficient? No, your phone just has a stupid DVFS table so the OEM can claim a high clockspeed in marketing materials and benchmarks. When run at the clockspeeds and voltages recommended by the SoC manufacturer you get much better battery life and temperatures. This is a valid criticism for shipping devices because nobody can expect users to root and swap DVFS tables, but it isn't on the OEM level because they can just NOT put in a stupid DVFS table and ship a more efficient device. Then again, it's true that Intel does power management vastly better...it's a shame they've been allergic to making non-garbage SoCs for the last few years (everything about old Atom was objectively wrong).

Nintendo Kid
Aug 4, 2011

by Smythe


Alereon posted:


But isn't ARM too slow? Too slow for what? It would be a long time before Apple could switch Macbook Pros (or iMacs or Mac Pros) to ARM, but their low-end devices like the Macbook Air and the now-retired Macbooks (and maybe even Mac Mini) are targets. ARM processors CAN scale up in performance, that is exactly what Apple has done with the Cyclone core in the A7 CPUs. More importantly, Apple would redesign their product stack around the strengths of the CPU lineup they actually had available to them.


ARM designs aren't putting out anything that can approach latest and second to latest generation Core i5s any time soon. And that's what's in the Airs and the Mini as the low end option.

It'd be a hell of a thing to "play up the strengths" to a level justifying how expensive those two products are in their base models while still having the real performance as low as they'd be.

Alereon
Feb 6, 2004

Dehumanize yourself and face to Trumpshed

College Slice

JawnV6 posted:

I think it's kinda stupid, but Heterogenous compute cores are one possible power/perf solution that isn't guaranteed or discarded yet. It seems pretty stupid to ship 40% more silicon just to give a chip a few more DVFS points to scale to, but there are shipping designs using it.
I feel like heterogenous compute (and particularly big.LITTLE) needs its own post. The problem is that there are multiple ways to do this, some are tremendously useful and some are worse than useless. ARM big.LITTLE works on "clusters" of up to four cores which share L2 cache. The obvious way to do this is to have two A7s and two A15s and shuffle tasks between them depending on the performance needed, this is called "core migration." The Samsung Exynos 5 Octa 54xx works on a method called "cluster migration", where it has a cluster of four A7s and a cluster of four A15s. Because there are two clusters there are two L2 cache modules, and there was supposed to be a Cache Coherency Interconnect between them to keep them in synch. This was actually broken on Exynos 5410 chips, so migrating between clusters has a huge performance penalty because data has to be flushed to main memory on migration. This was fixed on the Exynos 5 Octa 5420, but cluster migration is still a lazy technology unless you NEED four A15s in your phone (no you don't).

It seems to me that something like a single cluster of two A7s and two A12s provides much more useful dynamic range if you let the OS manage what cores are used and what tasks are assigned to those cores.

Alereon
Feb 6, 2004

Dehumanize yourself and face to Trumpshed

College Slice

Install Windows posted:

ARM designs aren't putting out anything that can approach latest and second to latest generation Core i5s any time soon. And that's what's in the Airs and the Mini as the low end option.

It'd be a hell of a thing to "play up the strengths" to a level justifying how expensive those two products are in their base models while still having the real performance as low as they'd be.
Intel thinks four Silvermont cores are a viable alternative to two Haswell cores for ~15W and below and low-end markets. Apple's Cyclone cores look like viable competitors to Silvermont at comparable TDP. Apple "redesigning their product stack" means just that: if they can't achieve Macbook Air performance (and thus price) with their SoC, call it something different or charge less. We're trying to fit ARM into Apple's existing Intel lineup, they're going to have a new lineup if/when they do this.

Henrik Zetterberg
Dec 7, 2007




Alereon posted:

Ivy Bridge supported something like Haswell's eDRAM and Apple was very excited about this (see how they bought all the R-series Haswells).

Do you have a source on this? As far as I was privy to, this isn't true. HSW (CRW really) was the first to support eDRAM.

e: Unless it was scrapped early in the design stage.

Henrik Zetterberg fucked around with this message at 19:15 on Nov 1, 2013

Nintendo Kid
Aug 4, 2011

by Smythe


Alereon posted:

Intel thinks four Silvermont cores are a viable alternative to two Haswell cores for ~15W and below and low-end markets. Apple's Cyclone cores look like viable competitors to Silvermont at comparable TDP. Apple "redesigning their product stack" means just that: if they can't achieve Macbook Air performance (and thus price) with their SoC, call it something different or charge less. We're trying to fit ARM into Apple's existing Intel lineup, they're going to have a new lineup if/when they do this.

Apple doesn't like to charge less ever. Their entire Mac business is based around not selling any laptops under $1000 and the Mini starts at $600 - though that's just for the computer and its power cord. They don't really make "low end" stuff, they just kinda accept they're never going to exceed 5% of the computer market ever again and stick to being a "high end" mark with mid to high end products.

And the slow ARM not-Macbook Air would find itself trying to compete with its kinda-OSX-but-not-really OS in a world where your average Windows laptop sells for around $400 and does have the ability to run everything people want. You might remember that this was same kind of "doesn't run any of your stuff you already have besides websites, but has great battery life" niche the original Linux-only netbooks tried to pull, only with nothing near the price advantage they had. And that those were quickly replaced by slightly more expensive, slightly lower battery life but full windows netbooks.

Alereon
Feb 6, 2004

Dehumanize yourself and face to Trumpshed

College Slice

Henrik Zetterberg posted:

Do you have a source on this? As far as I was privy to, this isn't true. HSW (CRW really) was the first to support eDRAM.

e: Unless it was scrapped early in the design stage.
Anandtech mentions it here, my understanding is that the L4 cache interface exists on the CPU but Intel never made the eDRAM chips to connect to it.

Alereon
Feb 6, 2004

Dehumanize yourself and face to Trumpshed

College Slice

Install Windows posted:

Apple doesn't like to charge less ever. Their entire Mac business is based around not selling any laptops under $1000 and the Mini starts at $600 - though that's just for the computer and its power cord. They don't really make "low end" stuff, they just kinda accept they're never going to exceed 5% of the computer market ever again and stick to being a "high end" mark with mid to high end products.

And the slow ARM not-Macbook Air would find itself trying to compete with its kinda-OSX-but-not-really OS in a world where your average Windows laptop sells for around $400 and does have the ability to run everything people want. You might remember that this was same kind of "doesn't run any of your stuff you already have besides websites, but has great battery life" niche the original Linux-only netbooks tried to pull, only with nothing near the price advantage they had. And that those were quickly replaced by slightly more expensive, slightly lower battery life but full windows netbooks.
You're going down this weird rabbit hole of worst-case scenarios. "What if Apple's SoC isn't good? What if they don't notice to late and the only way they have to compensate is by cutting price? What if it can't run OSX? What if they just switch to making lovely low-end Chromebook-ripoff devices how would they compete with Window?!" None of those are real scenarios. Apple's SoC will at least be competitive and Apple has a lot of different ways to tweak system performance to give users the kind of experience they want (such as their investments in fast PCI-E SSDs).

Nintendo Kid
Aug 4, 2011

by Smythe


Alereon posted:

You're going down this weird rabbit hole of worst-case scenarios. "What if Apple's SoC isn't good? What if they don't notice to late and the only way they have to compensate is by cutting price? What if it can't run OSX? What if they just switch to making lovely low-end Chromebook-ripoff devices how would they compete with Window?!" None of those are real scenarios. Apple's SoC will at least be competitive and Apple has a lot of different ways to tweak system performance to give users the kind of experience they want (such as their investments in fast PCI-E SSDs).

This isn't the worst case scenario, this is the best case scenario. There's nothing in switching to ARM for apple here that wouldn't be solved by simply having an official keyboard and touchpad for their iPads; and the Apple TV is already the ARM equivalent of what a Mac Mini ARM edition would be useful for.

And yes it wouldn't really be running OS X since real OS X will be sitting right there on their main computer products and is clearly going to get the lion's share of application support and importance.

An ARM version of a Macbook Air would be a lovely laptop that performs poorly and doesn't have the benefits of either the iOS or the OS X ecosystems, and would only have the dubious benefit of marginally better battery life. An ARM version of the Mac Mini would be a marginally more capable Apple TV, which you could already get if they simply upgraded the Apple TV again. These are best case scenarios, excepting the unlikely scenario of Apple switching all their stuff to ARM which would at least mean those things not being the third Apple OS ecosystem.

bull3964
Nov 18, 2000

DO YOU HEAR THAT? THAT'S THE SOUND OF ME PATTING MYSELF ON THE BACK.




The biggest issue I see is they aren't going to want to simultaneously run OSX on ARM and x86. They will either make a clean break to ARM for all their computers or fork the consumer OS into something else entirely.

WhyteRyce
Dec 30, 2001



The looming threat of Apple switching probably makes for a pretty good motivator to Intel and card to play in negotiations. Apple will end up getting what whey want in the end.

Nintendo Kid
Aug 4, 2011

by Smythe


bull3964 posted:

The biggest issue I see is they aren't going to want to simultaneously run OSX on ARM and x86. They will either make a clean break to ARM for all their computers or fork the consumer OS into something else entirely.

Consumer OS is all they have though. They killed their actual server business, and they've done plenty to drive away professional use as well as doing plenty to not encourage new use.

In essence, Apple is extremely ill-suited to being the people bringing a successful ARM switchover into being.

Alereon
Feb 6, 2004

Dehumanize yourself and face to Trumpshed

College Slice

Install Windows posted:

An ARM version of a Macbook Air would be a lovely laptop that performs poorly and doesn't have the benefits of either the iOS or the OS X ecosystems, and would only have the dubious benefit of marginally better battery life. An ARM version of the Mac Mini would be a marginally more capable Apple TV, which you could already get if they simply upgraded the Apple TV again. These are best case scenarios, excepting the unlikely scenario of Apple switching all their stuff to ARM which would at least mean those things not being the third Apple OS ecosystem.
I don't know why you keep repeating that ARM-based products are slow like it's a mantra. An ARM SoC in a cellphone or tablet is slow because it operates in a 1-4W TDP. Apple has designed a core that they can use to make a fast SoC in the ~10W+ space. Current products aren't overcoming the Turbo Boost advantage on i5s but we're not talking about current products.

bull3964 posted:

The biggest issue I see is they aren't going to want to simultaneously run OSX on ARM and x86. They will either make a clean break to ARM for all their computers or fork the consumer OS into something else entirely.
This is a very valid point, though Apple has supported multiple architectures simultaneously in recent history. It certainly wouldn't be fun for them and they would only do it if they thought it was worth it or had ways to make it much less troublesome.

Nintendo Kid
Aug 4, 2011

by Smythe


Alereon posted:

I don't know why you keep repeating that ARM-based products are slow like it's a mantra. An ARM SoC in a cellphone or tablet is slow because it operates in a 1-4W TDP. Apple has designed a core that they can use to make a fast SoC in the ~10W+ space. Current products aren't overcoming the Turbo Boost advantage on i5s but we're not talking about current products.
This is a very valid point, though Apple has supported multiple architectures simultaneously in recent history. It certainly wouldn't be fun for them and they would only do it if they thought it was worth it or had ways to make it much less troublesome.

There aren't any ARM SoCs performance-competitive with the i3s in cheap Windows laptops, let alone the i5s Apple has as their baseline in their computers. There won't be any anytime soon, because Intel is doing super well on continuing to run up performance and ARM designers at a whole aren't showing signs of being able to catch up fast enough. And again: on top of this Apple has already based their entire computer section on charging high prices and not selling anything low-end - putting in ARM replacements doesn't make sense there particularly when they already have the iPad line out there.

Apple never really supported multiple architectures. They have twice before had one old architecture being kept on life support and one new architecture that got all the new stuff. That's a plan for a transition of everything, not a plan or experience for actively maintaining two at once long term.

eames
May 9, 2009



Install Windows posted:

An ARM version of a Macbook Air would be a lovely laptop that performs poorly and doesn't have the benefits of either the iOS or the OS X ecosystems, and would only have the dubious benefit of marginally better battery life.

Warning, wild offtopic speculation ahead
I think the more likely route is a Macbook Air that can be separated at the hinge, with a fully functional but even lighter iPad Air as the display and a powerful CPU/Battery/IO/Keyboard for heavy work as the (optional) "base" part.

Itíll take a few years until we get there but that seems to be the obvious solution to me, although I have no idea how it would work on the software/OS side of things. I find it unlikely that they would mix two different architectures in one device, though.

JawnV6
Jul 4, 2004

So hot ...

Alereon posted:

It seems to me that something like a single cluster of two A7s and two A12s provides much more useful dynamic range if you let the OS manage what cores are used and what tasks are assigned to those cores.
The whole problem right now is that you're not letting the OS pick from any arbitrary config. It certainly would be nice to have an OS aware of the hardware and able to intelligently decide what to do, but we're nowhere close to that. Right now it's all bespoke hacks on top of ACPI to keep as much of the ugliness away from the OS as possible. And the end result without OS integration is half your compute silicon sitting idle bleeding current.

Alereon posted:

Why would Apple want to switch to ARM? Primarily so they can design and thus have total control over the SoCs in their own products. Intel has disappointed them on a number of occasions, for example Ivy Bridge supported something like Haswell's eDRAM and Apple was very excited about this (see how they bought all the R-series Haswells). Intel chose not to launch the eDRAM product and kind of left Apple in the lurch. Intel also still doesn't prioritize graphics (or drivers) in the way that Apple wants.
Just because they're disappointed in a missing feature doesn't magically grant them the capability to do it, and the entire surrounding system, themselves. Where's this in-house graphics team they've suddenly ginned up? They just going to buy Imagination too?

JawnV6 fucked around with this message at 18:56 on Nov 2, 2013

Nintendo Kid
Aug 4, 2011

by Smythe


eames posted:

Warning, wild offtopic speculation ahead
I think the more likely route is a Macbook Air that can be separated at the hinge, with a fully functional but even lighter iPad Air as the display and a powerful CPU/Battery/IO/Keyboard for heavy work as the (optional) "base" part.

Itíll take a few years until we get there but that seems to be the obvious solution to me, although I have no idea how it would work on the software/OS side of things. I find it unlikely that they would mix two different architectures in one device, though.

Honestly since iOS devices have already sold much more in the past 6 years they've been around than all Macs put together have sold since 1984, it's probably best for them to just continue focus on "make better iPads" rather than "attempt to cut up the small OS X market with cut down OS X computers on ARM".

One thing you could do with a docked iPad is simply provide it a way to cool down better, thus letting the CPU it already has sustain higher performance as long as it's in the dock. This could include simply putting the higher power CPUs in the iPad itself and having iOS simply refuse to run it in a high power state unless it detected it was in the appropriate dock.

(and yes I'm aware the article I linked is from last year, but since then the gap between iOS devices sold and Macs sold has gotten even huger)

Hendrik
Feb 5, 2009


Thank you everybody, very interesting responses.

eames posted:

Warning, wild offtopic speculation ahead
I think the more likely route is a Macbook Air that can be separated at the hinge, with a fully functional but even lighter iPad Air as the display and a powerful CPU/Battery/IO/Keyboard for heavy work as the (optional) "base" part.

Itíll take a few years until we get there but that seems to be the obvious solution to me, although I have no idea how it would work on the software/OS side of things. I find it unlikely that they would mix two different architectures in one device, though.

I wholly agree with this and that's what I meant by the netbook equivalent device, something that comes out of nowhere and becomes successful very rapidly. I think that there is a gap in Apples product range that they could fulfill with a Surface Pro equivalent device. They are in a good position with their iOS and Mac OS app stores that can be capitalized on by a low power consumption/productivity device. The problem with the iOS devices is that they are built for media consumption and any kind of productivity such as writing an email or a short word document is an exercise in frustration. My problem with the surface Pro is that it feels like a half-assed hybrid device with a unjustified premium price tag.

In conclusion the current Atom processors are not powerful enough and the i5 are too power hungry. A chip between the two could bring this proposed device into reality.

You Am I
May 20, 2001

I will decide who posts John Howard pictures in this thread and the circumstances in which they post.


I'm surprised Apple has the buying power to force Intel's hand at anything. Apple tried the same with IBM with the PowerPC CPUs and look how that went.

GrizzlyCow
May 30, 2011


Apple has gotten a lot bigger since then I do believe.

Anyway, they don't have to be the only ones pushing Intel to a solution. The market is moving more towards lower power and better iGPU which is what Apple wants too.

Nintendo Kid
Aug 4, 2011

by Smythe


GrizzlyCow posted:

Apple has gotten a lot bigger since then I do believe.

Not in the sector of the market where Intel CPUs are used though.

WhyteRyce
Dec 30, 2001



The Ultrabook was an attempt by Intel to save the PC market from itself and it just so happened that it basically is a Windows Air. Even if Apple isn't the biggest customer, it's probably a good idea to listen to one of the only partners who is selling high margin products that consumers universally love and is constantly pushing you to improve in new ways as opposed to some of the other guys who are racing to the bottom and messing up the Wintel experience (or further ruining depending on how you feel) by adding a layer of crap on top of everything.

WhyteRyce fucked around with this message at 03:55 on Nov 4, 2013

shrughes
Oct 11, 2008

(call/cc call/cc)


Install Windows posted:

Not in the sector of the market where Intel CPUs are used though.

Oh they certainly have.

Nintendo Kid
Aug 4, 2011

by Smythe


shrughes posted:

Oh they certainly have.

They most certainly have not. A bit under 5% of the computer market last year up from 3.5% of it the year they switched from PPC to Intel is not "a lot bigger". For comparison, Lenovo and HP were both pulling 15.5% with Dell and Acer both doing 10% last year.

And compare Apple's sub-5% in computers to its 25% in smartphones or 50% in tablets, and how both of those much more successful markets for them are all-ARM, so much so that last year they sold more iOS devices that year alone then Macs sold over all time. There's your impacts. Like seriously, somewhere around 190 million Macs sold since 1984 versus well over 200 million iOS devices in just 2012.

Nintendo Kid fucked around with this message at 03:57 on Nov 4, 2013

shrughes
Oct 11, 2008

(call/cc call/cc)


Apple's laptop market share is greater than 5%. Maybe you're thinking of the overall computer market share.

Nintendo Kid
Aug 4, 2011

by Smythe


shrughes posted:

Apple's laptop market share is greater than 5%. Maybe you're thinking of the overall computer market share.

They sold about 19 million laptops out of the 200 million laptops sold (the total computer market was about 310 million, so laptops were about 66%) so thats a bit under 10%. Other companies again sold much more. But again, they started off with about 7% of laptops sold in the year they switched from PPC to Intel, so that's not massive growth.

Again in comparison, Lenovo pulled in 16% share, the best they've ever had, HP was doing 15%, Dell was around 13% or so.

shrughes
Oct 11, 2008

(call/cc call/cc)


The size of the market grew like 300-400% though.

Nintendo Kid
Aug 4, 2011

by Smythe


shrughes posted:

The size of the market grew like 300-400% though.

And most other companies grew much more than them. Take Lenovo for instance, they've nearly tripled their share of the market from then to now.

Any way you slice it, Apple's only big growth has been in things that do not use Intel parts. Music players, smartphones, tablets, all on ARM.

eames
May 9, 2009



Intel open sourced their Broadwell GPU drivers. I expect they will be able to keep their +50% GPU performance per generation pace up.

quote:

The changes are massive and it's looking like the Broadwell graphics improvements will be astonishing and provide significant improvements over Haswell and earlier generations of Intel graphics.
While public details on Broadwell have been scarce beyond its 14nm fab process and new instruction set extensions, with the initial kernel driver code we have a feeling for Broadwell from the graphics side... Intel Broadwell graphics should be a terrific upgrade. Ben Widawsky in publishing the initial Broadwell support said, "Broadwell represents the next generation (GEN8) in Intel graphics processing hardware. Broadwell graphics bring some of the biggest changes we've seen on the execution and memory management side of the GPU. There are equally large and exciting changes for the userspace drivers."

Ben additionally said that the eigth-generation Broadwell graphics "dwarf any other silicon iteration during my tenure, and certainly can compete with the likes of the gen3->gen4 changes."

Canít wait to see some benchmarks of this chip compared to the respective Maxwell parts.

http://www.phoronix.com/scan.php?pa...ell_linux&num=2

Adbot
ADBOT LOVES YOU

JawnV6
Jul 4, 2004

So hot ...

eames posted:

Intel open sourced their Broadwell GPU drivers. I expect they will be able to keep their +50% GPU performance per generation pace up.

Uhh... did they also open source a few gigs worth of documentation required to touch any significant part of it?

  • 1
  • 2
  • 3
  • 4
  • 5
  • Post
  • Reply