The iPhone 5s Review
by Anand Lal Shimpi on September 17, 2013 9:01 PM EST- Posted in
- Smartphones
- Apple
- Mobile
- iPhone
- iPhone 5S
After Swift Comes Cyclone Oscar
I was fortunate enough to receive a tip last time that pointed me at some LLVM documentation calling out Apple’s Swift core by name. Scrubbing through those same docs, it seems like my leak has been plugged. Fortunately I came across a unique string looking at the iPhone 5s while it booted:
I can’t find any other references to Oscar online, in LLVM documentation or anywhere else of value. I also didn’t see Oscar references on prior iPhones, only on the 5s. I’d heard that this new core wasn’t called Swift, referencing just how different it was. Obviously Apple isn’t going to tell me what it’s called, so I’m going with Oscar unless someone tells me otherwise.
Oscar is a CPU core inside M7, Cyclone is the name of the Swift replacement.
Cyclone likely resembles a beefier Swift core (or at least Swift inspired) than a new design from the ground up. That means we’re likely talking about a 3-wide front end, and somewhere in the 5 - 7 range of execution ports. The design is likely also capable of out-of-order execution, given the performance levels we’ve been seeing.
Cyclone is a 64-bit ARMv8 core and not some Apple designed ISA. Cyclone manages to not only beat all other smartphone makers to ARMv8 but also key ARM server partners. I’ll talk about the whole 64-bit aspect of this next, but needless to say, this is a big deal.
The move to ARMv8 comes with some of its own performance enhancements. More registers, a cleaner ISA, improved SIMD extensions/performance as well as cryptographic acceleration are all on the menu for the new core.
Pipeline depth likely remains similar (maybe slightly longer) as frequencies haven’t gone up at all (1.3GHz). The A7 doesn’t feature support for any thermal driven CPU (or GPU) frequency boost.
The most visible change to Apple’s first ARMv8 core is a doubling of the L1 cache size: from 32KB/32KB (instruction/data) to 64KB/64KB. Along with this larger L1 cache comes an increase in access latency (from 2 clocks to 3 clocks from what I can tell), but the increase in hit rate likely makes up for the added latency. Such large L1 caches are quite common with AMD architectures, but unheard of in ultra mobile cores. A larger L1 cache will do a good job keeping the machine fed, implying a larger/more capable core.
The L2 cache remains unchanged in size at 1MB shared between both CPU cores. L2 access latency is improved tremendously with the new architecture. In some cases I measured L2 latency 1/2 that of what I saw with Swift.
The A7’s memory controller sees big improvements as well. I measured 20% lower main memory latency on the A7 compared to the A6. Branch prediction and memory prefetchers are both significantly better on the A7.
I noticed large increases in peak memory bandwidth on top of all of this. I used a combination of custom tools as well as publicly available benchmarks to confirm all of this. A quick look at Geekbench 3 (prior to the ARMv8 patch) gives a conservative estimate of memory bandwidth improvements:
| Geekbench 3.0.0 Memory Bandwidth Comparison (1 thread) | ||||||
| Stream Copy | Stream Scale | Stream Add | Stream Triad | |||
| Apple A7 1.3GHz | 5.24 GB/s | 5.21 GB/s | 5.74 GB/s | 5.71 GB/s | ||
| Apple A6 1.3GHz | 4.93 GB/s | 3.77 GB/s | 3.63 GB/s | 3.62 GB/s | ||
| A7 Advantage | 6% | 38% | 58% | 57% | ||
We see anywhere from a 6% improvement in memory bandwidth to nearly 60% running the same Stream code. I’m not entirely sure how Geekbench implemented Stream and whether or not we’re actually testing other execution paths in addition to (or instead of) memory bandwidth. One custom piece of code I used to measure memory bandwidth showed nearly a 2x increase in peak bandwidth. That may be overstating things a bit, but needless to say this new architecture has a vastly improved cache and memory interface.
Looking at low level Geekbench 3 results (again, prior to the ARMv8 patch), we get a good feel for just how much the CPU cores have improved.
| Geekbench 3.0.0 Compute Performance | ||||||
| Integer (ST) | Integer (MT) | FP (ST) | FP (MT) | |||
| Apple A7 1.3GHz | 1065 | 2095 | 983 | 1955 | ||
| Apple A6 1.3GHz | 750 | 1472 | 588 | 1165 | ||
| A7 Advantage | 42% | 42% | 67% | 67% | ||
Integer performance is up 44% on average, while floating point performance is up by 67%. Again this is without 64-bit or any other enhancements that go along with ARMv8. Memory bandwidth improves by 35% across all Geekbench tests. I confirmed with Apple that the A7 has a 64-bit wide memory interface, and we're likely talking about LPDDR3 memory this time around so there's probably some frequency uplift there as well.
The result is something Apple refers to as desktop-class CPU performance. I’ll get to evaluating those claims in a moment, but first, let’s talk about the other big part of the A7 story: the move to a 64-bit ISA.
Facebook
Twitter
RSS
464 Comments
View All Comments
MatthiasP - Tuesday, September 17, 2013 - link
Wow, first real review on the web AND deep as always, a very nice job from Anand. :)sfaerew - Wednesday, September 18, 2013 - link
Benchmarks(GFXBench 2.7,3DMark.Basemark X.etc.) are AArch64 version?There are 30~40% performance gap between v32geekbench and v64geekbench.
INT(ST)1471 vs 1065.
FP(ST)1339 vs 983
Wilco1 - Wednesday, September 18, 2013 - link
And Bay Trail Geekbench at 2.4GHz: 1063 (INT), 866 (FP)So A7 has beaten BT already by a huge margin despite BT not even being for sale yet...
TraderHorn - Wednesday, September 18, 2013 - link
You're comparing 64bit A7 vs 32bit BT. The 32bit #s are dead even. It'll be interesting to see if BT gets a similar performance boost when Win8 64bit versions are released in 1h 2014.Wilco1 - Wednesday, September 18, 2013 - link
BT's 32-bit result includes hardware accelerated AES, which skews its score (without it, its score is ~936). The 64-bit A7 result does also use hardware acceleration, so it is more comparable.Yes BT will get a speedup from 64-bit as well, but won't be nearly as much as A7 gets: its 32-bit result already has the AES acceleration, and x64 nearly isn't as different from x86 as A64 is from A32.
However the interesting things is that not even in 32-bit A7 wins by a good margin, but that it wins despite running at almost half the frequency of Bay Trail... Forget about Bay Trail, this is Haswell territory - the MacBook Air with the 15W 3.3GHz i7-4650U scores 3024 INT and 3003 FP.
Now imagine a quad core tablet/laptop version of the A7 running at 2GHz on TSMC 20nm next year.
smartypnt4 - Wednesday, September 18, 2013 - link
Why does the frequency matter? If the TDP of the chips are similar (Bay Trail was tested and verified by Anand as using 2.5W at the SoC level under load), who gives a flip about the frequency?If Apple wanted to double the frequency of the chip, they'd need something on the order of 4x the amount of power it already consumes (assuming a back-of-the-napkin quadratic relationship, which is approximately correct), putting it at ~6-8W or so at full load. That's assuming such a scaling could even be done, which is unlikely given that Apple built the thing to run at 1.3GHz max. You can't just say "oh, I want these to switch faster, so let's up the voltage." There's more that goes in to the ability to scale voltage than just the process node you're on.
Now, I will agree that this does prove that if Apple really wanted to, they could build something to compete with Haswell in terms of raw throughput. Next year's A8 or whatever probably will compete directly with Haswell in raw theoretical integer and FP throughput, if Apple manages to double performance again. That's not a given since they had to use ~50% more transistors to get a performance doubling from the A6 to the A7, and building a 1.5B transistor chip is nontrivial since yields are inversely proportional to the number of transistors you're using.
Next year will be really interesting, though. What with Apple's next stuff, Broadwell, the first A57 designs, Airmont, and whatever Qualcomm puts out (haven't seen anything on that, which is odd for Qualcomm.)
Wilco1 - Wednesday, September 18, 2013 - link
Frequency & process matters. Current phones use about 2W at max load without the screen (see recent Nexus 7 test), so the claimed 2.5W just for BT is way too much for a phone. That means (as you explained) it must run at a lower frequency and voltage to get into phones - my guess we won't see anything faster than the Z3740 with a max clock of 1.8GHz. Therefore the A7 will extend its lead even further.According to TSMC 20nm will give a 30% frequency boost at the same power. So I'd expect that a 2GHz A7 would be possible on 20nm using only 35% more power. That means the A7 would get 75% more performance at a small cost in power consumption. This is without adding any extra transistors.
Add some tweaks (like faster memory) and such a 2GHz A7 would be similar in performance as the 15W Haswell in MacBook Air. So my point is that with a die shrink and a slight increase in power they already have a Haswell competitor.
smartypnt4 - Wednesday, September 18, 2013 - link
Frequency and process matter in that they affect power consumption. If Intel can get Bay Trail to do 2.4GHz on something like 1.0V, then the power should be fine. Current Haswell stuff tops out its voltage around 1.1V or so in laptops (if memory serves), so that's not unreasonable.All of this assumes Geekbench is valid for comparing HSW on Win8 to ARMv8/Cyclone on iOS, which I have serious reservations about attempting to do.
The other issue I have is this: you're talking about a 50% clock boost giving a 100% increase in performance if we look at the Geekbench scores. That's simply not possible. Had you said "raise the clock to 1.6-1.7GHz and give it 4 cores," I'd be right behind you in a 2x theoretical performance increase. But a 50% clock boost will never yield a 100% increase with the same core, even if you change the memory controller.
Also, somehow your math doesn't add up for power... Are you hypothesizing that a 2GHz A7 (with 75% of the performance of Haswell 15W, not the same - as per Geekbench) can pull 2.6W while Haswell needs 15W to run that test? Granted, Haswell integrates things that the A7 doesn't. Namely, more advanced I/O (PCIe, SATA, USB, etc.), and the PCH. Using very fuzzy math, you can claim all of that uses 1/2 the power of the chip.
That brings Haswell's power for compute down to 7-8W, more or less. And you're going to tell me that Apple has figured out how to get 75% of the performance of a 7W part in 2.6W, and Intel hasn't? Both companies have ~100k employees. One is working on a ton of different stuff, and one makes processors, basically exclusively (SSDs and WiFi stuff too, but processors is their main drive). You're telling me that a (relatively) small cadre of guys at Apple have figured out how to do it, and Intel hasn't done it yet on a part that costs ~6x as much after trying to get deep into the mobile space for years. I find that very hard to believe.
Even with the 14nm shrink next year, you're talking about a 30% power savings for Intel's stuff. That brings the 15W total down to 10.5W, and the (again, super, ridiculously fuzzy) computing power to ~5-6W. On a full node smaller than what Apple has access to. And you're saying they'd hypothetically compete in throughput with a 2.6W part. I'm not sure I believe that.
Then again, I suppose theoretical bandwidth could be competitive. That's simply a factor of your peak IPC, not your average IPC while the device is running. I don't know enough about the low level architecture of the A7 (no one does), so I'll just leave it here I guess.
I'm gonna go now... I'm starting to reason in circles.
Wilco1 - Wednesday, September 18, 2013 - link
The sort of "simple" tweaks I was thinking of are: an improved memory controller and prefetcher, doubling of L2, larger branch predictor tables. Assuming a 30% gain due to those tweaks, the result is a 100% speedup at 2GHz (1.3 to 2.0 GHz is a 54% speedup, so you get 1.54 * 1.3 = 2.0x perf). The 30% gain due to tweaks is pure speculation of course, however NVidia claims 15-30% IPC gain for similar tweaks in Tegra 4i, so it's not entirely implausible. As you say a much simpler alternative would be just to double the cores, but then your single threaded performance is still well below that of Haswell.You can certainly argue some reduction in the 15W TDP of Haswell due to IO, however with Turbo it will try to use most of that 15W if it can (the Air goes up to 3.3GHz after all).
Yes I am saying that a relative newcomer like Apple can compete with Intel. Intel may be large, but they are not infallible, after all they made the P4, Itanium and Atom. A key reason AMD cited for moving into ARM servers was that designing an ARM CPU takes far less effort than an equivalent performing x86 one. So the ISA does still matter despite some claiming it no longer does.
smartypnt4 - Wednesday, September 18, 2013 - link
My point wasn't that Apple can't compete; far from it. If anything, the A7 shows they can compete for the most part. However, what you suggest is that Apple could theoretically have the same performance as Intel on a full node process larger at half the power. Ihave no illusions that Intel is infallible. Stuff like Larrabee and the underwhelming GPU in Bay Trail prove that they aren't. I just seriously doubt that Apple could beat Intel at its own game. Specifically, in CPU performance, which is an area it's dominated for years. It's possible, but I find it relatively unlikely, especially this early in Apple's lifetime as a chip designer.
On a different note, after looking at the Geekbench results more, I feel like it's improperly weighted. The massive performance improvement in AES and SHA encryption may be skewing the overall result... I need to dig more in to Geekbench before coming to an actual conclusion. I'm also still not convinced that comparing cross-platform results is actually valid. I'd like to believe it is, but I've always had reservations about it.