Sidebar Sidebar
Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.

What to expect from the iPhone 5S, Apple's A7 and beyond

chrmjenkins

chrmjenkins

macrumors 603
Original poster
Oct 29, 2007
5,325
158
MD
Time to update prediction from this:
A7
  • Manufacturer - Samsung on HKMG 28nm process
  • Die Size - 90-120 mm2
  • Designer - Apple
  • CPU Type - 1.3-1.6GHz Dual Core "Second Generation Swift Core"
  • Instruction Set - ARMv7s
  • Chip Designator - S5L8960X
  • L1 Cache - 32/32KB
  • L2 Cache - 1MB
  • RAM - 1GB LPDDR3 @ 1333 MHz (64 bit interface, PoP)
  • Max Theoretical Memory Bandwidth - 10.6 GB/s
  • GPU Type - "Quad Cluster" PowerVR 6430 @ 400 MHz
  • GPU Performance - 102.4 GFlops, 233 MTriangles/s

Bold is confirmed by Apple. Italics is changed from above prediction.

A7
  • Manufacturer - TSMC on HKMG HPM 28nm process
  • Die Size - 102 mm2
  • Transistors - approximately 1 billion
  • Designer - Apple
  • CPU Type - 1.5-1.8GHz Dual Core "Second Generation Swift 64-bit Core"
  • Instruction Set - ARMv8 with custom extensions
  • Chip Designator - S5L8960X
  • L1 Cache - 48/32KB
  • L2 Cache - 2MB
  • RAM - 1GB LPDDR3 @ 1333 MHz (64 bit interface, PoP)
  • Max Theoretical Memory Bandwidth - 10.6 GB/s
  • GPU Type - "Quad Cluster" PowerVR 6430 @ 270 MHz
  • GPU Performance - 69.2 GFlops, 157 MTriangles/s

See my explanation for these prediction changes over here: https://forums.macrumors.com/threads/1634100/

edit: the first GPU benchmark from 5S has been uploaded, supporting 2x claim.

Brian Klug noticed someone uploaded a graphics bench from 5S.

https://twitter.com/nerdtalker/status/377847764300099586
http://gfxbench.com/device.jsp?benchmark=gfx27&D=Apple+iPhone+5S&testgroup=overall

GFXBench 2.5 Egypt HD C24Z16 - Offscreen (1080p) : 56 FPS
GFXBench 2.5 Egypt HD C24Z16 - Onscreen : 53 FPS

Compared to iPhone 5:
GFXBench 2.5 Egypt HD C24Z16 - Offscreen (1080p) : 29.8 FPS
GFXBench 2.5 Egypt HD C24Z16 - Onscreen : 41.1 FPS
 
Last edited:
Y

yow.

macrumors member
Aug 26, 2012
52
0
That site also says: OpenGL ES 2.0 Apple A7 GPU - 27.10 (i.e. not OpenGL ES 3.0) http://gfxbench.com/device.jsp?benchmark=gfx27&D=Apple+iPhone+5S&testgroup=gl

I'm not sure what to make of that, since (IIRC) the apple presentation specifically said it supported ES 3.0.

But I think this point is crucial for whether this is a rogue, or uses the SGX554MP4. This was in the iPad 4, and gave double the graphics performance of the iPhone 5.

Note: I want it to be Rogue. But I'm playing devil's advocate because I've wanted it to be rogue on previous releases, and been disappointed. And I'm bitter.

My other point is that, even if it does support ES 3.0 after all, does that really rule out SGX554MP4 - perhaps with inhouse Apple modifications? Is it particularly hard to implement, so you might as well go for a whole new architecture?

*EDIT* apparently, yet another GPU - the SGX545 - can run openGL 3.2, and OpenGL ES 3.0 is a subset of that... so, they could be using it (though it would seem odd to make the lesser claim, unless perhaps they cut-down the functionality, to reduce power consumption): http://www.imgtec.com/news/Release/index.asp?NewsID=516 http://www.anandtech.com/show/6522/the-clover-trail-atom-z2760-review-acers-w510-tested/5
 
Last edited:
Galdom

Galdom

macrumors newbie
Mar 8, 2012
10
1
Sweden
yow. said:
That site also says: OpenGL ES 2.0 Apple A7 GPU - 27.10 (i.e. not OpenGL ES 3.0) http://gfxbench.com/device.jsp?benchmark=gfx27&D=Apple+iPhone+5S&testgroup=gl

I'm not sure what to make of that, since (IIRC) the apple presentation specifically said it supported ES 3.0.

But I think this point is crucial for whether this is a rogue, or uses the SGX554MP4. This was in the iPad 4, and gave double the graphics performance of the iPhone 5.

Note: I want it to be Rogue. But I'm playing devil's advocate because I've wanted it to be rogue on previous releases, and been disappointed. And I'm bitter.

My other point is that, even if it does support ES 3.0 after all, does that really rule out SGX554MP4 - perhaps with inhouse Apple modifications? Is it particularly hard to implement, so you might as well go for a whole new architecture?

*EDIT* apparently, yet another GPU - the SGX545 - can run openGL 3.2, and OpenGL ES 3.0 is a subset of that... so, they could be using it (though it would seem odd to make the lesser claim, unless perhaps they cut-down the functionality, to reduce power consumption): http://www.imgtec.com/news/Release/index.asp?NewsID=516 http://www.anandtech.com/show/6522/the-clover-trail-atom-z2760-review-acers-w510-tested/5
Click to expand...

Seems like if it is the SGX554MP4, wouldn't that be to power hungry for a phone?

Also got a question, would we see a higher frame rate if GFXBench was optimized for 64bit processor? Crunch numbers better perhaps? :)
 
chrmjenkins

chrmjenkins

macrumors 603
Original poster
Oct 29, 2007
5,325
158
MD
yow. said:
That site also says: OpenGL ES 2.0 Apple A7 GPU - 27.10 (i.e. not OpenGL ES 3.0) http://gfxbench.com/device.jsp?benchmark=gfx27&D=Apple+iPhone+5S&testgroup=gl

I'm not sure what to make of that, since (IIRC) the apple presentation specifically said it supported ES 3.0.

But I think this point is crucial for whether this is a rogue, or uses the SGX554MP4. This was in the iPad 4, and gave double the graphics performance of the iPhone 5.

Note: I want it to be Rogue. But I'm playing devil's advocate because I've wanted it to be rogue on previous releases, and been disappointed. And I'm bitter.

My other point is that, even if it does support ES 3.0 after all, does that really rule out SGX554MP4 - perhaps with inhouse Apple modifications? Is it particularly hard to implement, so you might as well go for a whole new architecture?

*EDIT* apparently, yet another GPU - the SGX545 - can run openGL 3.2, and OpenGL ES 3.0 is a subset of that... so, they could be using it (though it would seem odd to make the lesser claim, unless perhaps they cut-down the functionality, to reduce power consumption): http://www.imgtec.com/news/Release/index.asp?NewsID=516 http://www.anandtech.com/show/6522/the-clover-trail-atom-z2760-review-acers-w510-tested/5
Click to expand...

I think I can allay your fears a bit. A[] and A[]X processors have always had the same "sub-family" of GPU. A5 was 543MP2 and MP4. A6 was 543MP3 and A6X was 554MP4. 554 is more similar to 543 than the numbering suggests. First of all, 544 vs. 543 is simply API compliance differences. There's no meaningful compute difference. Second, the 554 is simply the 544 with double the ALU resources. That's why the FLOPs count doubles but the fillrate stays the same.

Now I'm going to make the argument of why A7 has to be Rogue because of A7X. Apple has typically at least doubled GPU performance each generation. To do that on A7X and keep 5XT, they'd either have to double the 554MP4 clock from 280 to 560 MHz, or double the core count to MP8. That kind of frequency is unheard of for 5XT and it's also incongruous with Apple's tradition of low clocks. MP8 is also unheard of and would simply be massive. A die space tradeoff would have had to occur where Rogue would have easily came out on top.

Rogue has been designed to scale easily in cluster size without significant front-end or back-end "glue" that complicated previous MP solutions.

Rest easy. If it's not Rogue, it's likely a custom Apple solution that dresses up the interfaces around standard Rogue execution paths.

Galdom said:
Seems like if it is the SGX554MP4, wouldn't that be to power hungry for a phone?

Also got a question, would we see a higher frame rate if GFXBench was optimized for 64bit processor? Crunch numbers better perhaps? :)
Click to expand...

Not meaningful differences. It's not likely CPU bound, so a couple FPS at most.
 
Galdom

Galdom

macrumors newbie
Mar 8, 2012
10
1
Sweden
From http://mashable.com/2013/09/12/iphone-5s-gaming/

The A7 chip has double the amount of registers compared to the A6. The amount of registers available to the applications can boost performance, especially for apps that do heavy computing, like high-end games," said Teppo Soininen, chief operating officer for Mountain Sheep, creators of Minigore. "The A7 chip is faster: 1.7GHz compared to the 1.3GHz A6. The A7 will most likely have a better memory bandwidth, and it will most likely be more energy efficient, which means you get more computing power for the same amount of power."
Click to expand...

Wonder if the true?:confused: 1.7GHz :rolleyes:
 
Y

yow.

macrumors member
Aug 26, 2012
52
0
chrmjenkins said:
MP8 is also unheard of and would simply be massive. A die space tradeoff would have had to occur where Rogue would have easily came out on top.
Click to expand...

Thanks! I was thinking of this, but it seems to hinge on whether Rogue is actually ready (and fulfills its amazing performance and efficient claims). MP8 seems plausible to me, since adding GPU cores scales well - at least, compared to CPUs. It wouldn't be the first time Apple did something unheard of!

However, i wasn't aware of the series 5 needing complicated glue for MP (though, if you can do MP4, MP8 doesn't sound that much more difficult). Secondly, Galdom's is right about power consumption, especially without a process shrink; but they are wizards... Actually, to be honest, I don't know how Apple managed to double CPU performance (maybe partly better cache/memory bandwidth?)

But to sum up: if Rogue *is* ready, performant and efficient, of course they'd use it (or base on it)! The kind of odd thing is that the GPU performance increase was so mild... hey... perhaps they aggressively underclocked the GPU, freeing up battery power for a higher clocked CPU? That would help explain the x2 CPU performance, and the rumored surprise jump to 1.7GHz...

Oh, BTW: I bet the M7 has another ARM core in it (low power and underclocked), making the 5S 3-core CPU; that's a way to effectively utilize extra CPU cores. A kinda-sorta big.LITTLE arrangement. Of course, it also might be special-purpose silicon... but harder to design, debug and make efficient; and less flexible, harder for programmers to make use of - and ARM + toolchain etc is already done.
 
You must log in or register to reply here.
Register on MacRumors! This sidebar will go away, and you'll see fewer ads.
Top Bottom