Apple's Swift: Pipeline Depth & Memory Latency

Section by Anand Shimpi

For the first time since the iPhone's introduction in 2007, Apple is shipping a smartphone with a CPU clock frequency greater than 1GHz. The Cortex A8 in the iPhone 3GS hit 600MHz, while the iPhone 4 took it to 800MHz. With the iPhone 4S, Apple chose to maintain the same 800MHz operating frequency as it moved to dual-Cortex A9s. Staying true to its namesake, Swift runs at a maximum frequency of 1.3GHz as implemented in the iPhone 5's A6 SoC. Note that it's quite likely the 4th generation iPad will implement an even higher clocked version (1.5GHz being an obvious target).

Clock speed alone doesn't tell us everything we need to know about performance. Deeper pipelines can easily boost clock speed but come with steep penalties for mispredicted branches. ARM's Cortex A8 featured a 13 stage pipeline, while the Cortex A9 moved down to only 8 stages while maintining similar clock speeds. Reducing pipeline depth without sacrificing clock speed contributed greatly to the Cortex A9's tangible increase in performance. The Cortex A15 moves to a fairly deep 15 stage pipeline, while Krait is a bit more conservative at 11 stages. Intel's Atom has the deepest pipeline (ironically enough) at 16 stages.

To find out where Swift falls in all of this I wrote two different codepaths. The first featured an easily predictable branch that should almost always be taken. The second codepath featured a fairly unpredictable branch. Branch predictors work by looking at branch history - branches with predictable history should be, well, easy to predict while the opposite is true for branches with a more varied past. This time I measured latency in clocks for the main code loop:

Branch Prediction Code
  Apple A3 (Cortex A8 @ 600MHz Apple A5 (2 x Cortex A9 @ 800MHz Apple A6 (2 x Swift @ 1300MHz
Easy Branch 14 clocks 9 clocks 12 clocks
Hard Branch 70 clocks 48 clocks 73 clocks

The hard branch involves more compares and some division (I'm basically branching on odd vs. even values of an incremented variable) so the loop takes much longer to execute, but note the dramatic increase in cycle count between the Cortex A9 and Swift/Cortex A8. If I'm understanding this data correctly it looks like the mispredict penalty for Swift is around 50% longer than for ARM's Cortex A9, and very close to the Cortex A8. Based on this data I would peg Swift's pipeline depth at around 12 stages, very similar to Qualcomm's Krait and just shy of ARM's Cortex A8.

Note that despite the significant increase in pipeline depth Apple appears to have been able to keep IPC, at worst, constant (remember back to our scaled Geekbench scores - Swift never lost to a 1.3GHz Cortex A9). The obvious explanation there is a significant improvement in branch prediction accuracy, which any good chip designer would focus on when increasing pipeline depth like this. Very good work on Apple's part.

The remaining aspect of Swift that we have yet to quantify is memory latency. From our iPhone 5 performance preview we already know there's a tremendous increase in memory bandwidth to the CPU cores, but as the external memory interface remains at 64-bits wide all of the changes must be internal to the cache and memory controllers. I went back to Nirdhar's iOS test vehicle and wrote some new code, this time to access a large data array whose size I could vary. I created an array of a finite size and added numbers stored in the array. I increased the array size and measured the relationship between array size and code latency. With enough data points I should get a good idea of cache and memory latency for Swift compared to Apple's implementation of the Cortex A8 and A9.

At relatively small data structure sizes Swift appears to be a bit quicker than the Cortex A8/A9, but there's near convergence around 4 - 16KB. Take a look at what happens once we grow beyond the 32KB L1 data cache of these chips. Swift manages around half the latency for running this code as the Cortex A9 (the Cortex A8 has a 256KB L2 cache so its latency shoots up much sooner). Even at very large array sizes Swift's latency is improved substantially. Note that this data is substantiated by all of the other iOS memory benchmarks we've seen. A quick look at Geekbench's memory and stream tests show huge improvements in bandwidth utilization:

Couple the dedicated load/store port with a much lower latency memory subsystem and you get 2.5 - 3.2x the memory performance of the iPhone 4S. It's the changes to the memory subsystem that really enable Swift's performance.

 

Apple's Swift: Visualized Six Generations of iPhones: Performance Compared
POST A COMMENT

278 Comments

View All Comments

  • grkhetan - Wednesday, October 17, 2012 - link

    Multiple display reviews conclude that the iPhone 5 has the best display in a smartphone (And much better than a Samsung Galaxy S 3)

    http://www.displaymate.com/Smartphone_ShootOut_2.h...

    http://www.anandtech.com/show/6334/iphone-5-screen...
    Reply
  • rarson - Thursday, October 18, 2012 - link

    Your second link doesn't compare it to anything but the iPhone 4. Your first link ONLY compares it to the S3. Neither link supports your statement ("best display in a smartphone"). Reply
  • doobydoo - Friday, October 19, 2012 - link

    A quote from his second link:

    'To put this in perspective, in the past few years I've reviewed probably 30-40 different displays, from PC monitors to TVs to projectors. Not a single one, out of the box, can put up the Gretag Macbeth dE numbers that the iPhone can, and perhaps one projector (which listed for $20,000) can approach the grayscale and color accuracy out of the box.'
    Reply
  • steven75 - Wednesday, October 17, 2012 - link

    Those pesky facts are annoying! Reply
  • Obsoleet - Tuesday, October 16, 2012 - link

    No, it's not. There's many reasons the GS3 is the better choice based on the software and hardware, mainly that the MaxxHD only matches a 5 month old phone in hardware specs and tosses on a bigger battery as the only clear win (but you get stuck with a Motorola phone vs most people's preferred choice Samsung).
    But the killer reason is that the charger is on the left hand side.

    For many of us lefties, that is a deal breaker. As a right handed user, you don't realize this. I want the ports on the top or bottom, and I just ordered a GS3 because of this being a tipping point.

    The original Maxx had the ports on the top! Motorola is clueless.
    Never again.
    Reply
  • Ckaka1993 - Thursday, December 06, 2012 - link

    Ppi does make a difference. Go see the videos of droid dna(has 440) ppi and you can make out the difference. iPhone 5 doesnt have true 720p but that doesn't matter cause it's quite close to 720p. Anyways iphone5 is behind so many smartphones at present. Nokia lumia 920 is a treat to watch with its 332 ppi pure motion hd+ display and high refresh rate, u can make out the difference. But nexus 4 is the smartphone which gives u the best worth for money at ony 350usd it is freaking awesome Reply
  • makken - Tuesday, October 16, 2012 - link

    The Physical Comparison table lists the iPhone 5's resolution at 1136 x 960, instead of 1136 x 640. Threw me off for a second there =P Reply
  • Brian Klug - Tuesday, October 16, 2012 - link

    Oops, fixing. There's always something in the table that needs fixing it seems :P

    -Brian
    Reply
  • DukeN - Tuesday, October 16, 2012 - link

    And always favorably on the Apple side.

    Maybe you took a picture of the pixel count with the iPhone's camera...
    Reply
  • Alucard291 - Tuesday, October 16, 2012 - link

    I know I love how their own benchmarks show how the battery life is worse in just about everything than the 4s and yet and yet "its better" >.> Reply

Log in

Don't have an account? Sign up now