Apple dumps Intel from laptop lines
- Thread starter rpg.314
- Start date
May be not i7, but I am sure in that time frame ARM cores will be out which can rival mid range CPU's in perf.
Seriously, assuming 64 bit, OoO, 3 issue with 2 way MT on top, I see no reason why ARM can't give intel a run for it's money in the mid range.
I meant SoC vendors copying that and putting dedicated hw in their chips.
Blazkowicz
Legend
does the future ARM have virtualization extensions? not only a Mac question, they want to get used more in servers too.
..self answer, yes it's there from cortex A15 on, and it's old news.
http://www.arm.com/products/processors/cortex-a/cortex-a15.php
can someone point to that famous article on ARMv8?
rpg., is that triple issue core with SMT a v8? a bit confused as I can't help but think an instruction set is not an architecture.
there's encoding hardware and some image processing hardware in nvidia Tegra already. Intel copied the SoCs rather than the other way around. it was long overdue on PC, I remember encoding was vaporware on the NV40 which is a good while ago now.
..self answer, yes it's there from cortex A15 on, and it's old news.
http://www.arm.com/products/processors/cortex-a/cortex-a15.php
can someone point to that famous article on ARMv8?
rpg., is that triple issue core with SMT a v8? a bit confused as I can't help but think an instruction set is not an architecture.
there's encoding hardware and some image processing hardware in nvidia Tegra already. Intel copied the SoCs rather than the other way around. it was long overdue on PC, I remember encoding was vaporware on the NV40 which is a good while ago now.
I was speculating about the stock core licensed by ARM. Architectural licensees might do better.
Blazkowicz
Legend
even the current core i3 2100T makes an impressive showing, and it's relatively cheap.
Apple can also request the best and most undervolted Intel bins. here the Core2duo SU series already kick Atom and Bobcat in the nuts.
if ARM reaches that performance level at lower power use it will be pretty impressive.
Frankly, I know much more people with personal non-work MacBooks than with Windows based laptops, Silicon Valley Reality Distortion Field and all that. And my colleagues who use a MacBook at work also use it in OSX mode. I don't know any of them who runs Bootcamp. Some have Parallels or Vmware, just in case, but nobody really uses it.
That said, I'm sure such users exist.
Don't all current ARM SOCs come with HW video decoder and encoder? Doesn't that qualify as a transcoding solution? It's probably much better performance than quick sync (assuming the particular format is HW supported.)Then you mention quick sync - an Intel solution to bolster my argument? Thanks!Now show me the ARM solution with a dedicated transcoding solution...
I think it's not unthinkable that Apple will switch to ARM in the long term. Apple is great at vertical integration of high margin parts. The CPU would definitely qualify.
Personally, I wouldn't mind, as long as the performance is there (==equivalent to, say, current MacBook Air). Ever since I switched to MS Office for Mac, my Vmware license has been gathering dust.
Bingo, although quite a few do share some hardware between encode & decode so their transcoding performance isn't always as impressive as you might expect. However, that's not true of the Imaginations VXE for example, and from everything I know I'm confident that actually has higher video quality than Intel Quick Sync.
VXE also scales up in performance through both clock speeds and the number of internal pipelines, which is customer configurable. With 4 pipes instead of the standard 2 and a clock speed of 400MHz instead of the standard 200MHz (which is probably very realistic on 28nm High-K), you could transcode *to* 1080p H.264 40Mbps High Profile at 120fps. Well, you could if you had enough memory bandwidth that is, and if the VXD decoder also scaled up performance as easily which it doesn't. Still, 2x to 4x real-time for 1080p depending on the bitrate seems realistic to me.
So yeah, that's not ARM's job (excluding their own hardware video IP which is on a short hiatus while they work on the next-gen architecture) and it's really not a point in Intel's favour at all if the SoC manufacturer does their job well.
1.6GHz Core 2 Duo? Oh man, you could beat that silly with a 28nm Dual-Core Cortex-A15 on 28HP. And if you're willing to sacrifice a little bit of single-threaded performance in exchange for higher multi-threaded one, quad-core A15 vs 2.4GHz Core 2 Duo in the current Macbook is the same thing.
The problem, of course, is that: 1) The Macbook Pro has a quad-core i7 which is quite something else. 2) CPU performance will continue to increase. So I rather doubt you'd like Macbook Air performance 2.5 years from now... On the other hand, if you had even half the Macbook Pro's CPU power budget, you could optimise your synthesis and your binning a lot more towards performance and you could still come up with something very impressive (8-core 3.5GHz on 20HP?)
---
Practically, I don't think any of this discussion matters because:
1) It's not impossible at all that Apple is still working on a higher performance in-house ARM core. On the right process and optimised sufficiently for performance (as opposed to area and possibly power), that could be more than good enough. We're not talking about a 250mW CPU here.
2) If Apple switches to ARM for 'laptops', I don't think it's to come up with 'yet another laptop'. Capacitive touch screen and the major OS update to go along with it is a minimum in my mind. As opposed to a tablet, laptops will always be defined as being larger and with a physical keyboard. Everything else is an implementation detail.
2GHz arm cores and what will pcs be at ?
Cpus have decreased in power usage while increasing performance on the x86 side and the trend continues look at sandy bridge vs the first gen i series and ivory bridge seems to be bringing more power decreases with performance increases.
It seems silly to me. Mabye if its something between the ipad and laptop (net book ?)
I haven't used the latest MacBook air, but the first one is way too slow.
I mainly use parallels on my bootcamp partition to test PowerPoint translation between Mac/PC, to run a few win-only tools and to intall new steam purchases to my bootcamp partition. I also use it to run two virtual servers and to play with things like Android and webOS.
Just a side gripe...as pretty as Office 2011 for OS X is, it is every bit as bug ridden as its win counterpart. OpenOffice is more stable IMHO. I also have to laugh when someone needs help troubleshooting Outlook on OS X; why are they running that pos?
I mainly use parallels on my bootcamp partition to test PowerPoint translation between Mac/PC, to run a few win-only tools and to intall new steam purchases to my bootcamp partition. I also use it to run two virtual servers and to play with things like Android and webOS.
Just a side gripe...as pretty as Office 2011 for OS X is, it is every bit as bug ridden as its win counterpart. OpenOffice is more stable IMHO. I also have to laugh when someone needs help troubleshooting Outlook on OS X; why are they running that pos?
The main bugs are importing graphics and then moving/resizing them. If the graphics are one-step from the origin you're all good, but when it's a graphic from an old version CorelDraw from 12 years ago or a converted .eps from a whie back that has been converted between formats, etc. PowerPoint will often vomit upon any attempt to resize, etc. whereas OpenOffice had no issues with the same files.
As a result I have recreated many of my long-standing reference graphs that I've used for background in much of my research.
Then there's the fact that the same native Powerpoint file will display slightly differently between various versions even within the same OS platform. I often have an equation, or chemical reaction on a single line that winds up split in some bizarre fashion on the conference laptop. This would happen when using 2007 and the conference had 2003.
It's even worse with 2011->2010 (osx to win) as the fonts are slightly different and sometimes you get random weird characters in plots for some reason.
As such I try to always use my own laptop at conferences. When this isn't possible I "touch up" in 2007 and hope that's what the conference is using.
Do you have video/sound/fancy transitions in your presentations? If not, export to PDF is the safe way to go.
Cheers
I find PDF doesn't present well...artifacting of graphics if the resolution of the projector is off, irregular pagination (PgDwn) in Reader, etc. but I have done it at times.
I don't do any fancy prancy stuff
Ditching x86 for mid and high end laptops is highly unlikely in the near to mid term. Apple has come a long way from the 'over priced, under performing junk'-stigma of their 68K/PPC product lines to the aura they have today. They changed their image by consistently producing high quality and high performance parts (still overpriced though). There is no way they are going back.
I could imagine low end notebooks switch to an ARM SOC. Effectively an iPad 2+ with a keyboard. Possibly running iOS. If that takes off, I could see them try to move that up market.
Cheers
I could imagine low end notebooks switch to an ARM SOC. Effectively an iPad 2+ with a keyboard. Possibly running iOS. If that takes off, I could see them try to move that up market.
Cheers
I could see this "rumor" ending up being true. Apple doesn't want people using Windows programs on their Macs. They want people to use Mac programs designed for Macs. Also as long as the CPU is powerful enough to run the most common Mac Apps then I don't see the need for excessive CPU hardware. Also since they're already using ARM processors for their phones and tablets it's probably better to unify all of their software to use the same ISA not to mention cheaper prices on CPUs. Since Windows will support ARM it makes even more sense as porting would be extremely easy. I could see future Macs using ARM cpus and PowerVR graphics chips.
I think you will see cheaper Mac books with long battery life when they start to use ARM processors. Think netbook but more powerful yet low cost.
I think you will see cheaper Mac books with long battery life when they start to use ARM processors. Think netbook but more powerful yet low cost.
Last edited by a moderator:
Are there any benchmarks available for the faster ARM chips vs desktop chips from Intel?
The problem with an Apple netbook is, it's basically overlapped with iPad.
If Apple wants to compete in netbook territory, what their netbook will be? There are at least three possibilities:
1. a Macbook with Atom (i.e. traditional netbook)
2. a Macbook with ARM, running iOS
3. a Macbook with ARM, running MacOS X
the first possibility is boring, but not impossible. The second one is also boring, because it's just an iPad with a keyboard, which is already available now (you can connect a keyboard to an iPad already). The third one is interesting and the only relevant one to this discussion, but I'd say it's nearly impossible at least for near term.
The reason is simple. How big is the market of netbook? Right now, the cheapest Macbook is at US$1000 price level, while iPad is at US$500 price level. If Apple makes an ARM based netbook, it's very likely the price level will be between the two. However, there are already many good notebooks (not netbook, real notebook) in this price level. So basically Apple will have to compete with real notebooks with their netbook, which is already a relatively small market, and with limited software support. I just don't see how this will work
From current development trend I think it's more likely that Apple will somehow provide the ability to run iOS applications on a Macbook, either through emulation, fat binary, or an additional processor module. This way, if they want to dump x86 all together in the future, it will be easier.
If Apple wants to compete in netbook territory, what their netbook will be? There are at least three possibilities:
1. a Macbook with Atom (i.e. traditional netbook)
2. a Macbook with ARM, running iOS
3. a Macbook with ARM, running MacOS X
the first possibility is boring, but not impossible. The second one is also boring, because it's just an iPad with a keyboard, which is already available now (you can connect a keyboard to an iPad already). The third one is interesting and the only relevant one to this discussion, but I'd say it's nearly impossible at least for near term.
The reason is simple. How big is the market of netbook? Right now, the cheapest Macbook is at US$1000 price level, while iPad is at US$500 price level. If Apple makes an ARM based netbook, it's very likely the price level will be between the two. However, there are already many good notebooks (not netbook, real notebook) in this price level. So basically Apple will have to compete with real notebooks with their netbook, which is already a relatively small market, and with limited software support. I just don't see how this will work
From current development trend I think it's more likely that Apple will somehow provide the ability to run iOS applications on a Macbook, either through emulation, fat binary, or an additional processor module. This way, if they want to dump x86 all together in the future, it will be easier.
Similar threads
