alt plus = gives this
≠
That is not equals to.
We Mac users don't have to use substitute symbols like =! when the actual symbol is so easy to use ≠.
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.
Apple Reportedly Tapping TSMC for 20-nm Quad-Core Chips with Late 2013 Production
- Thread starter MacRumors
- Start date
- Sort by reaction score
alt plus = gives this
≠
That is not equals to.
We Mac users don't have to use substitute symbols like =! when the actual symbol is so easy to use ≠.
They'd have to come out with a deal that included licensing agreements from both AMD and Intel, likely VIA too. I don't see that happening.
I can think of three reasons why you won't:
1. Tis' the season; exactly a year ago people were talking about iPad 2 DH
2. Apple kept the 30 pin connector on iPhone 4S, they could have switched it if they have an overnight migration strategy, but they didn't, which means they will continue to sell at least one product with the 30 pin connector for two more years
3. we haven't seen a single leaked part for the iPad 3 with Lightning switch.
Did I miss something? Other than speculation, have we seen a proof?
Why doesn't apple buy TSMC?
TSMC is a very different business than Apple's. Lots of customers might leave TSMC if treated the way Apple operates, which would quickly destroy the value of the investment.
What Apple could buy, is a fab for TSMC. They cost many many billions. Let TSMC build and run it (there appear to be only 4 or 5 companies in the world that know how to do this well), and Apple gets some absolutely guaranteed production output. TSMC gets a "free" fab after meeting their commitment (in both volume and quality).
And the fact that Apple isn't offering to do this with Intel, just down the road, who also can build world class fabs, may suggest something about Apple's future plans for processors.
I'm not convinced that anyone actually can license the x86 ISA. AMD, Cyrix and VIA all seem to be running on legacy licenses, and the x86-64 would have to be a AMD/Intel license combo and that would probably close to impossible to facilitate. There's not a lot of companies doing x86 processors, and I think there's licensing issues in the way for this. I have no other idea why the likes of TSMC, IBM, Freescale, Broadcom, Qualcomm, Ti and Samsung (and a multitude of Chinese companies) wouldn't build x86 processors of their own.
The i486 architecture is older than 20 years and patens are probably expired. i686 (Pentium Pro) came in 1995 so these patents would probably expire soon too.
I can't see the point though. Apple ship A LOT more products using ARM and there are A LOT more software for Apple's iOS platform than OSX, and this gap is only expanding. A lot of code is in Cocoa, C or some other more or less easily portable code. So, It's more likely that Apple will license/design a high performance (with high frequency and strong floating point, out of order operation, 64-bit, branch prediction, wide and fast external buses, PCI, etc) ARM architecture to replace its x86 offerings than to make an x86 microarchitecture on its own.
LIKE AN APPLE: Microsoft will design their own ARM chip based on an ARM Mali graphics core. WinRT will run on a dedicated Microsoft Windows Processor (No more Qualcomm, Nvidia fragmentation). TBA '13.
Since TSMCs market cap is 78 billion, I think you might be wrong. However, I see no reason why Apple would want to buy them.
----------
You know thats not dollars right?
ARM+own GPU in SoC for MacBook and iMac? Maybe a 24-core(6-chips) in a MacPro in blade configuration is not far off. What saddens me is Apple's clear intention to move away from x86 architecture to make their products unique marks the end of Hackintosh users for there won't be any compatible motherboards for us to run OSX any longer.
Unless Apple really wants to finally kill off their Mac line or they have a secret engineering team that can out engineer Intel's best, they won't be doing this anytime soon. Even the fastest ARM designs have nothing on a Core 2 Duo circa 2006.
Not true.
http://en.wikipedia.org/wiki/TSMC
But it will cripple apple (100B ish total price?)They dont need to buy a fab company.
Haha, this is some funny stuff. I stopped reading AppleInsider forums, because I couldn't stand how you derailed conversations with your absolutely useless remarks, Tallest Skil. I came here to avoid having to read your nonsense, and now look where you show up.
I'd ask you to stop being so snide, but that would be pointless. Instead I'll implore everyone else on this forum to keep up the quality discourse, and ignore your trolling.
I know I'm falling for your trap here, but I can't stand idly by and watch another website I really enjoy get corrupted by you.
If you need evidence of my point: http://forums.appleinsider.com/t/153327/apple-likely-to-unveil-ipad-mini-at-event-on-oct-23-report/80#post_2210241
Last edited:
Why doesn't apple buy TSMC
Because TSMC made it clear they are "not for sale"? The article said both Apple and Qualcomm both attempted bids to buy exclusive rights to TSMC. That presumes that Apple and Qualcomm were willing to pay a LOT of money to lock in TSMC. But apparently TSMC does not want to be locked in by either one. If they won't go for that, then TSMC certainly won't go for the idea of selling themselves to anyone (yet). In plain English, that means that TSMC wants to retain its independence (for now).
true, Apple doesn't need to or want to buy TSMC. Actually it would cost at least $112B assuming the average 43% acquisition premium.
Hahahaha
Good one. You sir are wrong. I am not Tallest Skil. He is was banned while I was a user here. We used to share 'tars. At any point, your prejudice should be used against you instead.
That's not it. Both Apple and Qualcomm both wanted TSMC to set aside production capacity for them and neither wanted to buy TSMC.
Besides it's unlikely Apple is interested in buying TSMC. Apple wants to stay away from managing manufacturing back ends. It's like saying Apple should buy LG Display or Foxconn. I don't think Apple is interested in that.
It's the very thing Tim Cook moved Apple away from. There is no reason to believe he is looking to reverse course.
From 32 nm to 28 to 20 nm. I love the speed increase and power use reductions this implies. iPads will take over more MacBook functions as long as clever app producers are motivated. Maybe iPads will be able to use more physical devices wirelessly to help this along.
Only because the fastest ARM cores, before today, have been designed by smaller companies for a much lower power envelope.
With MSWord getting ported to the ARM, hundreds of thousands more iOS apps than Mac apps, more workload getting farmed to the "cloud", data centers looking to lower their total energy costs, number crunching moving to parallel arrays and GPUs, etc.; legacy x86 cores may start getting pushed out of the consumer and enterprise mainstream, squeezed between ARM on one hand, and data centers plus new Big-Iron on the other. x86 could get pushed into a niche, just as the once dominant 360/370 ISA was. Intel CEOs should be paranoid.
Mostly because many of us are lazy to do that. Also, in programing, != is used as not equal too.
Of course you could build a chip based on ARM with greater power and power draw. It would have nothing in common with today's ARM processors other than the instruction set, so anything about it is pure speculation.
There is no evidence that anybody can do better processors in the desktop/laptop power sweet spot than Intel just because they're big enough. AMD sure is trying. Instructions sets don't matter much for that chip size, because the instructions are decoded anyway (and the decoders are small compared to the rest of the chip). What matters is experience (+1 for Intel) and structure size (+1 for Intel).
That's not what they did with the A6. It's a custom design, but its not made from scratch. It uses a lot of bits (ha ha) and pieces from the excellent ARM designs. Going from there they could beat the reference designs.
Apple can't beat Intel by building an x64 CPU from scratch. It's too huge a task, and there is no base to start from. And Intel is always a structure shrink ahead. Remember, IBM couldn't beat Intel with their own architecture, let alone on their home turf.
Somehow I recall that Intel only fabs for itself. So that'd simply not be an option.
----------
theBB's correct about TSMC and fab incompatibilities.
And you're correct that adding suppliers is a good idea... which is what this speculation implies as well..
A5/A6 chips have to be manufactured by Samsung.
But next gen Ax chips manufactured by TSMC avoids disrupting the Samsung/Austin fabs.
Now you have two suppliers covering your production. Sounds good to me.
