Yes they do. The A7 is noticeably faster in many apps, not just gaming. iOS 7 stutters a bit on the iPhone 5, no stutter on the 5S (in multitasking) - but this is probably due to lack of optimization.
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.
Inside Apple's A7 Chip, M7 Motion Coprocessor, and More from the iPhone 5s
- Thread starter MacRumors
- Start date
- Sort by reaction score
Yes they do. The A7 is noticeably faster in many apps, not just gaming. iOS 7 stutters a bit on the iPhone 5, no stutter on the 5S (in multitasking) - but this is probably due to lack of optimization.
I disagree. A7/M7/64-bit technology is right here, right now.
The ball is at the hands of the iOS apps developers.
The next six / twelve months will be exciting times for iPhone 5s users.
Next year will just be going thru the motion. Another cycle of pissing contests between the fanboys.
The ball is at the hands of the iOS apps developers.
The next six / twelve months will be exciting times for iPhone 5s users.
Next year will just be going thru the motion. Another cycle of pissing contests between the fanboys.
This...
which means take data from many sensors, and compact it down to basically a delimited data stream of variables directly actionable by apps.
It's an impressive SoC to be sure, plenty of different functional blocks inside. It would be wonderful to see them all labelled and the bus topology added.
Behold, the chip of tomorrow, today!
https://www.macrumors.com/2013/09/2...a7-processor-2-4-times-faster-for-some-tasks/
Its the majority of the 5S. One guy tested many in the store and they all did it. There is a really good thread with lots of proof/photos but a moderator closed it because some people were getting a bit nasty and derailing the thread. No pattern or color.
I'm just hoping it helps conserve battery while I'm looking at the moving home screen.
Why wait for a descendant?
I ported some of my extremely compute intensive engineering calculations from my 2009 MacBook Pro to iOS, recompiled for arm64 using the new Xcode, and found that the calculations now benchmark just as fast on an iPhone 5s as on the MacBook Pro (which BTW is a lot faster than was a desktop G5 "supercomputer").
No waiting, equal performance, better battery life, and, very unlike the G5, it fits in my pocket.
Cool article. I've submitted an app that utilizes the M7 motion chip to compute step counts in the background as you walk with your iPhone without using any processing power (at least not from the main A7 chip). I think we should see a lot interesting uses of this new technology soon.
There is a very large block of SRAM/cache above the GPU (looks around 4MB) that contributes positively to the density. That contributes 192 million transistors on its own. Add the 1MB of L2 from the CPU and you're at 240 million, almost a quarter of the total transistor count.
edit: Here's anandtech's annotated image. The CPU cores are outlined in yellow. The four GPU clusters and supporting blocks are in blue.
http://www.anandtech.com/show/7355/chipworks-provides-first-apple-a7-die-shot
adding cache to the CPU increases performance more than increasing MHz or GHz. this has been known for decades
the fandroids and bloggers seem to look only at paper specs
Did you try OpenCL calculations to see how they would work on the 5s GPU's? That would be pretty amazing in itself.
A pocket analysis tool with a range of optional BTLE and Lightning sensors; what would somebody to with something like that, Bones?
It's not that easy to make a qualitative statement like that. The performance boost of cache depends heavily on your type of workload. Increasing cache size also increases its latency. It also matters how associative your cache is. It also depends how much frequency increase you're talking about. For reference, Apple did not increase the L2 cache on the A7 from what the A6 (or A5) had.
Not if the apps you use happen to fit in the smaller cache. Too big a cache for your apps can actually slow a processor down by adding to load latencies.
If an app just happens to completely fit in the bigger cache, but severely trashes the smaller cache, then yes there can be a huge (sometimes order of a magnitude or more) speed-up in performance of that app from the added cache.
No significant real world benefit? I think you´re reading the wrong reviews m8.
The A7 is almost as fast as my i7 in web browsing performance, it´s a LOT faster in video processing (as fast as a 2009 Macbook Pro), it´s extremely powerful for games, it´s 8x as fast when doing AES encryption (for VPN use in companies this can improve throughput) and there are so many other things like a faster PowerVR encoder/decoding IP core (which still wasn´t even mentioned, let alone analysed by tech sites, seems I might be doing my own review soon for that).
Only thing I miss is WiFi 802.11ac and LTE advanced (Germany already has it) support, but I can live with that. It´s already blazing fast.
And if you re-engineer or re-compile some of your CPU-intensive applications for the A7, there´s also a benefit that varies according to how CPU- or GPU-bound you really are. Developers need to play catch-up right now to even come close to the potential of the A7.
Which is the reason why the A7 might easily be the chip with the most longetivity of every Ax-chip, yet.
Last edited:
From what I've seen, rotation animation speed is independent of processor. It might be stutter-ier with a slower processor but will generally complete a 90 degree rotation in a more or less fixed amount time, because it's an animation, not a task.
Have you tried recompiling apps for arm64 and benchmarking the difference? I'm an app developer, and I have with my apps (engineering apps, not games), and you are SO wrong. There's a clear difference on many apps that are not games.
For other non-gaming uses, look at the Javascript benchmark for the 5s. Every complicated web page (that's lots of major sites these days) gets a noticeable performance benefit from the A7.
Realistically, how far away are we from Apple being able to produce their own chips for all of their devices? 10 years?