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's iPad Chip Development Cost Estimated in $1 Billion Range
- Thread starter MacRumors
- Start date
- Sort by reaction score
that is why
This is why we love Apple and tolerate all the others me to companies.
Its not like anyone gets excited or even cares about say Dell or Denovo.
Sure we may have to use them because the company we work for makes us do it, but is anyone really excited about anything new most of these companies come out with.
There are a few companies I keep an eye for cool products, and no MS$ is not one of them. I just tolerate the unfinished OS because of the gaming side, but for work and about everything else its my trusted Macbook Pro.
This is why we love Apple and tolerate all the others me to companies.
Its not like anyone gets excited or even cares about say Dell or Denovo.
Sure we may have to use them because the company we work for makes us do it, but is anyone really excited about anything new most of these companies come out with.
There are a few companies I keep an eye for cool products, and no MS$ is not one of them. I just tolerate the unfinished OS because of the gaming side, but for work and about everything else its my trusted Macbook Pro.
In the wider, real, world - not many people are excited when
Apple puts chips that the other companies have been using for
months into a new Apple box.
We like our Core i5/i7 laptops - too bad that Apple's kit is so
far out of date.
Don't know about Nvidia or Qualcomm, but Apple very likely owns an ARM architecture license (which probably costs millions), which would allow them to design an ARM CPU implementation from scratch. Marvell has an architecture license (originally belonging to Dobberphul's group at DEC, then to Intel), and has a CPU very different (except for the ARM ISA of course) from the Cortex line. Motorola (now Freescale) may have also purchased an ARM architecture license, but doesn't appear to have done anything with it.
Apple is worth 181 Billion with 40 Billion in the bank
Lets put this in perspective!
Based on the their stock price, market capitalization, Apple is worth 181 Billion, yes even more than Google. According to their SEC audited financial reports, Apple has 39 Billion in the bank.
Also, Microsoft is only worth 80 billion more than Apple at $251 Billion.
So 1 Billion for a data center in NC and another Billion to build their own chip is not that much money for Apple!
Think Bigger!
Marcus
Lets put this in perspective!
Based on the their stock price, market capitalization, Apple is worth 181 Billion, yes even more than Google. According to their SEC audited financial reports, Apple has 39 Billion in the bank.
Also, Microsoft is only worth 80 billion more than Apple at $251 Billion.
So 1 Billion for a data center in NC and another Billion to build their own chip is not that much money for Apple!
Think Bigger!
Marcus
A part of me feels Schadenfreude, as Microsoft would never be able to switch so easily from one processor platform to another as Apple did so numerous times thanks to UNIX I suppose. 
Who would have thought that the general trend does not go to more power, but towards more flexibility and MS cant cope with this.
Who would have thought that the general trend does not go to more power, but towards more flexibility and MS cant cope with this.
Why on Earth would it cost that much? It's not even using an original CPU design-just taking ARM's Cortex A9 core(s?). The only design here is slapping all the parts together. Not trivial of course...but a billion dollars? That seems insane.
I assume it's using more PowerVR graphics too?
I assume it's using more PowerVR graphics too?
The NYT article doesn't actually say it cost that much. The NYT article cites no sources explaining where the 1 billion figure comes from. It actually costs less than $100 million.
Can we stop talking about this now?
A part of me feels Schadenfreude, as Microsoft would never be able to switch so easily from one processor platform to another as Apple did so numerous times thanks to UNIX I suppose.
Who would have thought that the general trend does not go to more power, but towards more flexibility and MS cant cope with this.
Never?
Windows 7 client runs on two processor platforms - 32-bit x86 and 64-bit x64.
Windows 7 "server" runs on two processor platforms - 64-bit x64 and 64-bit IA64.
Windows Mobile/CE runs on x86 and compatibles, MIPS, ARM, and Hitachi SuperH processors.
Previous versions of Windows NT-based systems have been sold and supported on PPC and Alpha processors. SPARC and MIPS were also running (but not officially sold).
I agree except for the Win7 stuff. As a long-time x86 microprocessor designer who was one of the team who invented AMD64 let me tell ya - there's almost no work to be done to support it. That was how we destroyed Intel's dopy Itanium nonsense.
Personally, I saw NT running on PPC back in the day when I was designing PPC processors, and it blew away Mac OS in speed and reliability at the time.
Of course, one difference is that Apple, when it switches architectures, makes it so most programs run on any architecture (either using fat binaries or emulation). Don't remember if NT ever did that. At one time MS's mobile stuff had some way to do that if I recall from my PocketPC programming days.
Itanium was some expensive *****! We had some production SQL Servers on HP IA64, and the happiest work day of my life was the day we successfully load tested the x64 platform and found it had comparable performance at a fraction of the cost.
Crap, there was no compatibility at all! Microsoft at the time wouldn't even develop SQL admin tools for their own platform. Seriously, not even their own dtsrun.exe executable was ported to IA64!!! WTF!?! LOL!
BTW, the second happiest work day of my life was the day we put those IA64 boxes on pallets and shipped them off for someone to repurpose them as HP/UX or Alpha or whatever other HP platform they might be "useful" on. I was just glad to be rid of them and hope I never see another one.
/threadjack
I remember waaaaaay earlier in the decade where I was wondering if it was dumb for me to buy a new computer at the time because probably within a year we'd have to be switching to Itanium :-D
SOOOOOO glad trusty old x86 won out. The funny thing is, had Intel not been trying to shove Itanium down our throats, THEY would have gotten to define 64-bit x86. I wonder if they'd have done a better job at it...that's always bugged me. Like AMD only bumped externally addressable registers from 8 to 16, and claimed x86 didn't need more, while PowerPC chips (back when they made those ) had 32. I'm sure there's other stuff that *might* be questionable. Oh well, serves Intel right.
SOOOOOO glad trusty old x86 won out. The funny thing is, had Intel not been trying to shove Itanium down our throats, THEY would have gotten to define 64-bit x86. I wonder if they'd have done a better job at it...that's always bugged me. Like AMD only bumped externally addressable registers from 8 to 16, and claimed x86 didn't need more, while PowerPC chips (back when they made those
) had 32. I'm sure there's other stuff that *might* be questionable. Oh well, serves Intel right.
I remember waaaaaay earlier in the decade where I was wondering if it was dumb for me to buy a new computer at the time because probably within a year we'd have to be switching to Itanium :-D
SOOOOOO glad trusty old x86 won out. The funny thing is, had Intel not been trying to shove Itanium down our throats, THEY would have gotten to define 64-bit x86. I wonder if they'd have done a better job at it...that's always bugged me. Like AMD only bumped externally addressable registers from 8 to 16, and claimed x86 didn't need more, while PowerPC chips (back when they made those
We did 16 because the tradeoff when you go to 32 is that task switching takes much longer (because you have to backup a bigger register file), and going from 16 to 32 had little net benefit for most real code.
x64 binaries are not compatible with x86, therefore it's a unique architecture compiled with a unique compiler code generator.
x64 also requires 64-bit clean code - all the "compatibility" in
the world won't help if you store a pointer in a 32-bit int.
______________
The point is that Windows NT is mostly architecture-neutral code,
and Microsoft has been shipping it on quite a few platforms. The
fact that x64 is similar (though not a superset) of x86 doesn't
mean much to OS developers writing in high level languages.
A part of me feels Schadenfreude, as Microsoft would never be able to switch so easily from one processor platform to another as Apple did so numerous times thanks to UNIX I suppose.
Who would have thought that the general trend does not go to more power, but towards more flexibility and MS cant cope with this.
Hogwash. Windows NT 4.0 supported four processor architectures: x86, Alpha, MIPS and PowerPC.
x86-64 is binary compatible except for long mode 64-bit apps (i.e.: in 4 out of 5 modes). And it's a lot easier to change a define for a type than it is to switch to a fundamentally different ISA. At AMD we worked with Cutler on 64-bit WoW, and their architecture made it much harder than it needed to be; for Linux, to build 64-bit versions we literally had to do nothing other than change a compiler flag (of course, we had to hack our own compiler back in the day
True. And it was the best OS for PowerPC (and the worst for Alpha
I heard that there was also OS/2 for PowerPC.
To the C++ programmer, using a compiler with a different code
generator is beyond simple.
Changing the type definitions, however, is work.
____________
To the high level programmer, changing from x86 to Alpha is the
same effort as changing from x86 to x64 (that is, a compiler
change).
The initial 64-bit Windows port was done on Alpha,
and IA-64 was out a couple of years before x64.
The biggest advantage to x64 for the developer/end-user is that
the similarity to x86 means that silicon can run 32-bit code at
nearly the same speed as 64-bit code.
Exactly. And now that they put their own silicon into some of their devices, the profit they would have given to someone else by purchasing their silicon is just a pure cream. Clearly there will be a desire to expand the iPhone OS product line, as I would expect these products to be quite profitable for Apple. I suspect the investment in the development of A4 will be recouped rather quickly.