Actually any app compiled with the current iOS SDK will be 64-bit. So many popular apps are already running on the A7's 64 bit execution path. The speed advantages will be due to the differences in armv8 though, not due to 64-bit addressing.
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.
Qualcomm Marketing Executive Calls Apple's 64-Bit A7 Chip a 'Marketing Gimmick'
- Thread starter MacRumors
- Start date
- Sort by reaction score
Actually any app compiled with the current iOS SDK will be 64-bit. So many popular apps are already running on the A7's 64 bit execution path. The speed advantages will be due to the differences in armv8 though, not due to 64-bit addressing.
so are you haters.
could you not be bothered to read Mike Ash's very long, very detailed post about just that?
RAM has little to do with speed (on the phones) - apart from the situations where Safari has to reload pages in multi-tab browsing because it does not have enough memory to keep all the pages in-memory. Also, the statement that iOS devices are faster is patently incorrect. iPhone 5S beats, say, Galaxy Note 3 in some benchmarks (mostly those that use single thread) and loses in others (it will lose in all tasks that scale well with multithreading)
We agree, and I didn't suggest that they shouldn't. I did marketing for six years. Of course it should appear in there, especially given that it's a loaded term that people associate with being better. That said, when it's included in the same paragraph as performance improvements, it's a bit misleading becauseas we've already establishedit's all that other (admittedly complex junk that 99% of people don't understand and therefore isn't good in marketing collateral) stuff that is creating those improvements.
Ultimately, I think that's all the Qualcomm guy was trying to highlight, which is why in my original statement I said he's factually correct. (Just to be 100% clear here, a lot of marketing is inherently a bit misleading, and I don't necessarily have a problem with that, except when it's entirely disingenuous. Back in my original statement, all i said was, "Apple's rhetoric doesn't match up with the evidence.") I think we agree that you and I were having a semantic argument that's largely irrelevant. Of course the A7 is a significantly faster chip. Of course that's due to improvements, some of which were made easier due to the move to 64-bit architecture. I don't think even the dude at Qualcomm would argue otherwise.
It is much less of an issue for android, as most apps executed in virtual machine they don't care about processor architecture, only android itself has to be ported and this should not be difficult. From other hand android devices usually have more ram that ios counterparts so I expect to see 64bit processors there soon enough - most likely in a first half of the next year.
Did you bother reading the next sentence as to why and what/who will benefit from it? If your going to quote at least get it all.
Yep, I'll grant you that, but those calculation improvements are relatively small compared to the other architectural improvements that are made easier with 64-bit architecture, which is where the gains are derived.
Nope, not insinuating that at all, per my other posts in this thread.
My issue is that Apple has nowhere said it's better or faster only because it's 64 bit. They provided independent speedup metrics.
Yet, people respond to the fact that it's 64-bit as if they is what they are saying, or they are faulting Apple for people making that incorrect assumption. The plain fact is that they would have been silly to not mention it or suggest it won't be faster for any reason. They're criticizing apple for marketing it exactly as they would have, if not more disingenuous. For instance, they love to talk about core counts, whereas apple didn't even bother this time.
Thus, it all comes across to me as "I'm mad you got to talk about this as a feature before I did."
Exactly! But why would you stick with 32 when you'll need to go 64 anyway in the future?
And all the thousands of apps would have to be updated by that time, so you might as well start earlier, right?
That's not really the point. Just because it makes little or no difference today, it is nearly certain that it will make a significant difference in years to come. and the minor performance improvements already out there make it worth it to release the chip now. It is a chicken and egg situation - the chip needs to come before anything will be optimized for it.
also it won't help you check you email, thats for sure. its hard to comment on which part is responsible for the crazy performance gains. needless to say, the competition is a bit shaken up
Well it can't be a marketing gimmick if the vast majority of people have no clue about it, and most people don't even know what 64 bit means and don't care. Apple aren't even advertising this as a feature, they simply mentioned it.
And since we will need it when the time comes, when 4GB of RAM becomes a reality, every app will have long been written to take full advantage of that, rather than having to rewrite code at the last minute.
And since we will need it when the time comes, when 4GB of RAM becomes a reality, every app will have long been written to take full advantage of that, rather than having to rewrite code at the last minute.
It's not just the hardware improvements. The ARMv8 64-bit instruction set has also been cleaned up and optimized. Instruction pipelining and execution are faster than for the (still fully supported) 32-bit instruction set.
Also, Apple has been using "tagged pointers" in OS X since 10.7. This allows storing small amounts of data within 64-bit pointers themselves, instead of allocating memory then dereferencing pointers to that memory to get and set values. This is a huge win for instances of the NSNumber object, for example. And now that's available in iOS as well as OS X.
Another huge win is storing object reference counts in objects themselves instead of in a separate table (that requires locking for thread safety.) This speeds up ARC-based Cocoa memory management, which is critical to iOS and its apps. And that in-object reference count wouldn't be possible with 32-bit ints and pointers.
These optimizations and others are all enabled transparently, from the developer's perspective, by Xcode. So no, it's not just the silicon that's been upgraded to 64-bit. It's the OS, runtime, apps, frameworks, etc. And all of it, together, results in massive performance gains.
