Separate names with a comma.
Discussion in 'CPUs and Overclocking' started by grimpr, Nov 5, 2012.
Guess Charlie had a point, even with all the rage against him. Anyway, the merge of iOS and OSX is inevitable. Intel still has time for a major turnaround in their mobile offerings.
This really isn't much of a surprise, and Apple is right, one day they will be powerful enough.
I think the real interesting question is what timeframe are they planning on having this done, and what are they going to do about the fact that by that time ARM CPUs are powerful enough to comfortably run laptops, there will likely be x86-based portables that look just as nice, and get 'good enough' (anything longer than 12hrs is just gravy) battery life that the extra performance of x86 is preferable?
That depends. Eventually they will be as powerful as today's x86, yes (probably in around a decade), but will today's hardware be enough to run the applications of 2022? That is by no means a given.
Bleeding edge has always seemed very fast, but year after year, the bar gets raised as well. What we felt was crazy fast in 2002 is laughable now (and is actually where ARM currently is).
I think the really big question is, "how much money and time will it take to stay a few years behind bleeding edge?" Today, Core 2 would be the performance to shoot for. In 2-3 years, Nehalem. And so on. I wonder, if you plotted R&D cost v. performance at any given time (get out yer 3D graphing software! ), how many years you'd have to stay back to remain at or under the knee on that curve?
ARM themselves are pulled in too many directions. But, I could see a company like nVidia, Samsung, or Apple developing ARM CPUs far enough behind bleeding edge to not break them, but not a decade behind PCs, ether. Around 5 years behind would be enough for the majority of the consumer market.
Well, considering where Apple is going with their desktop apps, and how quickly they are adding performance to their ARM ISA parts, it's no stretch of the imagination to think that in four or five years a custom Apple designed ARM cpu will be enough to handle the all the needs of Apple's leisure computing devices.
That won't happen anytime in the near future simply because ARM is just so much slower than even ulv intel offerings. While Apple would probably be the only company that could get away with selling crappy hardware at a high price here the difference is just too big. Especially in the productivity sector they would never ever get away with it. I mean your 3 year old Mac will be double the speed as a new one. No one in their right mind would buy new for years.
The trade off being that the battery would run for about 10 days instead of 2 hours.
I'd love to see Adobe CS6 run on current ARM chips.
Which really isn't the point of the OP's linked article or this thread.
Apple either underestimates the Intel juggernaut or is truly ignorant. Nobody is catching up, not even the Chinese. This has to be a strategic (almost coercive) maneuver to eke out higher profit margins from suppliers like Intel.
ARM ISA isn't going to become a replacement for Intel, not across the board. But if you look at what Apple is becoming, which is a provider of a platform for leisure computing, then they won't need to focus on providing products with 100+Gflop CPU's.
IMO, what Apple is wondering is what will it take to make ARM ISA a viable replacement for their laptops and desktop parts, since their focus clearly isn't on high performance computing anymore.
Any gamer will tell you, the CPU is very much secondary to the GPU in gaming performance. I imagine that if the goal is something for watching movies, web browsing, light gaming, etc., the CPU only has to be fast to run the OS and GUI smoothly and that's pretty much it. As long as the GPU they put in it is fast, I don't think anyone would notice much of any downgrade in performance.
Same rumours, same reruns, no products.
We heard this since 2010 or maybe even before. AMD Apple will use APUs, Apple will use ARM. Nothing never happens.
Sure about that? The guy is a professional liar.
What i found ironic about this is - i'm sure whomever is designing ARM SOCs and custom cores within Apple wants this to happen.
More profits, same ISA - it's natural.
What i find ironic is - it will pretty much kill off they're original resurgance market of "creative professionals".
Which means all the hipsters that made them a popular quality lined product will be forced to leave them - as they become what... Acer?
Which i'll put my nutsack down on happening if apple ever switches.
They can't sustain the whole brand image - if they drop the 1 inside component that makes sure the experience they're trying to sell is seamless to the consumer.
I kinda hope it happens for apple's own dooms sake.
ARM A15 is still lightyears behind any intel desktop offering, even old Core Duos.
ARM has a long way to go before they can topple x86 for the desktop
Pretty much, if they want to be able to compete against Windows 8.
this is old news. obviously they would like to make their own chips like they do for ipad but arm performance is so far off from intel that they couldnt afford to do this in the near future. you cant just come out with something significantly lower performance and sell it as an upgrade
Ofcourse Apple is exploring ways to replace Intel. Seems kinda obvious. What's interesting is that they need to go public with it.
Probably just a bargaining chip with Intel for a better price.
Apple: "We want a better price or we are switching to a competitor!"
Intel: "What competitors?"
Windows 8 isn't even a worthy competitor. There is no unanimous agreement that W8 is the best successor to W7. If iOS and Mac OSX were to merge the way that W8 did, it will suffer the same backlash.
The google chromebook review showed the A15 to be faster then the old atom one. A custom apple A15 would be even quicker. I wouldn't be surprised if the macbook air started using one of them. Apple would save a lot of money and gain a lot of flexibility if they could dump Intel/x86.
As for desktop - do apple even have desktop pc's these days?
That's also my guess. The question here is if Intel can eat a discount large enough to keep Apple motivated not to develop a competing product.
Apple could use 10 years and 25billion$ and not even reach Intel speeds.
And Apple can never sell enough on its own to be profitable that way. Not to mention Apple is having a continuel pressure froms tock owners to pay up, specially after the 120$ stockloss the last 50 days.
^. apple would still be limited by TSMC/Samsung no? unless they go into the fab game but to even come close to intel would cost 10s of billions just to build the fabs i think. let alone the R&D to do it
best i can think of is that (afaik apple was in on the ground floor of open CL?) they want to have more run on the gpu maybe via directcompute or some such? if enough can be run that way, then even desktops could run ok (for mac users i mean) with a decent gpu and maybe 2 quad A15s?