The A9X SoC & More To Come

Finally, as everyone is undoubtedly eagerly anticipating our look at the A9X SoC inside the iPad Pro, let’s take a very quick look at what we know about the SoC so far. There’s a bit of a limit to what we can do blindly via just software, but I’m hoping that the eventual A9X die shots will confirm some of our suspicions on A9X’s configuration.

Apple SoC Comparison
  A9X A9 A8X A6X
CPU 2x Twister 2x Twister 3x Typhoon 2x Swift
CPU Clockspeed 2.26GHz 1.85GHz 1.5GHz 1.3GHz
GPU PVR 10 cluster Series7? PVR GT7600 Apple/PVR GXA6850 PVR SGX554 MP4
Memory Bus Width 128-bit 64-bit 128-bit 128-bit
Memory Bandwidth 51.2GB/sec 25.6GB/sec 25.6GB/sec 17.1GB/sec
L2 Cache 3MB 3MB 2MB 1MB
Manufacturing Process Unknown
(TSMC 16nm or Samsung 14nm)
TSMC 16nm &
Samsung 14nm
TSMC 20nm Samsung 32nm

First and foremost, the most unexpected news here is that unlike A8X, A9X is not packing a triple-core CPU. Instead A9X drops back down to just a pair of Twister CPU cores. The twist here is that relative to A8X and A9, Apple has cranked up their CPU clockspeeds. Way, way up. Whereas the iPad Air 2 (A8X) shipped at 1.5GHz and the iPhone 6s (A9) at 1.85GHz, the A9X sees Apple push their clockspeed to 2.26GHz. Not counting the architectural changes, this is 22% higher clocked than the A9 and 51% higher than the A8X.

The fact that Apple dropped back down to 2 CPU cores is unexpected given that we don’t expect Apple to ever go backwards in such a fashion, and while we’ll never know the official reason for everything Apple does, in retrospect I’m starting to think that A8X was an anomaly and Apple didn’t really want a tri-core CPU in the first place. A8X came at a time where Apple was bound by TSMC’s 20nm process and couldn’t drive up their clockspeeds without vastly increasing power consumption, so a third core was a far more power effective option.

WebXPRT 2013 (Chrome/Safari/IE)

Basemark OS II 2.0 - Overall

By comparison, with the FinFET process Apple is using here – and given the lower volume of A9X I don’t have reason to believe it’s dual-sourced, so it’s either TSMC or Samsung – Apple has been free to increase their clockspeeds substantially. At the same time these FinFET processes are still new and yields won’t be great, so there is a strong incentive to keep die sizes down to keep yields up, and adding a third core would only make that harder. If I had to guess, Apple only wanted two cores to begin with – this makes it easier for developers knowing that they only have two cores to work with – and that it’s A8X that is the anomaly.

Otherwise a highly clocked CPU is far more in-line with Apple’s design philosophy as it means that A9X is capable of amazing single-threaded performance – and keep in mind that we’re talking ARM Cortex-A57-like clockspeeds for a CPU that gets much more work done per cycle – so what we see here makes a lot of sense. Plus with iPad Pro in particular Apple has more battery capacity to sustain the power draw of a higher clocked SoC, and more surface area to dissipate that heat, so the usual concerns about power and cooling aren’t quite as pressing. I do wonder if this will impact multitasking performance much, but given what Twister is capable of, I’m not nearly ready to write off a dual-core Twister implementation clocked this high.

Moving on, as is customary for the X-series SoCs from Apple, A9X features what I believe to be a wider 128-bit LPDDR4 memory bus. The memory bandwidth numbers clearly point to a wider bus, and Apple needs the bandwidth to feed a more powerful GPU.

Geekbench 3 Memory Bandwidth Comparison (1 thread)
  Stream Copy Stream Scale Stream Add Stream Triad
Apple A9X 2.26GHz 20.8 GB/s 15.0 GB/s 15.3 GB/s 15.1 GB/s
Apple A8X 1.5GHz 14.2 GB/s 7.44 GB/s 7.54 GB/s 7.49 GB/s
A9X Advantage 46.4% 101% 103% 102%

Which brings us to the last bit of our preview, the GPU. Apple went with a 6 cluster PowerVR Series7 design on A9, and for A9X they have gone with a larger design. Without a die photo it’s basically impossible to determine how many clusters are in use since clockspeed plays such an important role. What we do know is that GPU performance relative to A9 has pretty much doubled, which once again is right in-line with Apple’s usual design goals.

GFXBench 3.0 Manhattan (Offscreen)

Given what Apple has done with clockspeed on Twister, for the moment I am staking my bet on it being a 10 cluster design with a higher GPU clockspeed than A9 giving us the rest of the performance boost. To be clear here this could also be a 12 cluster design at a similar clockspeed or even an 8 cluster design clocked far higher – we’ll need die shots to confirm – but given all of the options it’s a 10 cluster design that is the best balance between die size and clockspeed, and it would also be the biggest curveball Apple could throw. It should also be noted that PowerVR Series7 certainly supports such a configuration since it’s scalable from 2 to 16 clusters, although in Imagination’s official product catalog they don’t have a name for such a configuration. So for the moment I’m simply calling it a 10 cluster Series7.

Anyhow, we’ll be back later with a full review of the iPad Pro, including the pros and cons of Apple’s first large-format, productivity-oriented tablet, and a full breakdown of the A9X SoC. So until then stay tuned.

Taking Notes With iPad Pro
Comments Locked


View All Comments

  • Spunjji - Wednesday, November 11, 2015 - link

    100% agreed with that sentiment... always pushing the new.
  • HollyDOL - Wednesday, November 11, 2015 - link

    Ouch, that equation brings back memories...
  • SaolDan - Wednesday, November 11, 2015 - link

    I would like to see what that GPU could do in a x86 game. I guess we will never know:(
  • id4andrei - Wednesday, November 11, 2015 - link

    Not much. All rendering in mobile devices is done at half precision. Imagination's GPU is by design excellent for it.
  • tipoo - Wednesday, November 11, 2015 - link

    That's how the scores with only 50GB/s memory bandwidth for the whole chip together (not just GPU alone) are so high I guess, doesn't need as much bandwidth at half precision. Are these still tile based renderers? That saves bandwidth too. That's how you have this chip getting close to the Iris Pro/650M, if you don't take into account it's using half precision.
  • id4andrei - Wednesday, November 11, 2015 - link

    Precisely. Also from Anandtech, in a past article analyzing PowerVR, the tiling arch is especially adept at FP16 workloads.

    Macrumors are going insane with the Ars GFX benches. They actually think that A9x GPU(great as it may be) is faster than the one in the 13" rMBP(28W SoC).
  • akdj - Tuesday, December 29, 2015 - link

    I'm I'd group Macrumors 'as one', nor the conversation you're referring to without spirited debate (about Iris Pro and Imagination's 10 core implementation in the iPP)
    They're certainly not 'going insane' ...but there's certainly something to the raw bench speeds on graphic oriented applications, their usage of the resources and most importantly, the software or as we call it now, the apps developed specific to that SoC versus a Skylake 6xxx with an iGPU attempting to run X86 software that is a lot better off with discrete GPU(s). IMHO, that's where the 'win' for ARM comes in, versus Intel
    It's not ALL about benchmarking but weird enough; sites have been using the same comparisons since the earliest days of Android and iOS --- not a single complaint!
    Apple drops the A7, the first 'mobile' SoC with 64bit capability sucking the wind out of Qualcomm execs and kicking Intel in the rear when it comes to better iGPU integration with their mobile bound laptop or tablet chips.
    The A8's increase overall in computing and graphic power ...a marked improvement in a single year, a year Qualcomm and others STILL had a helluva time getting a working 64bit processor to market and the A8x, along with Swift and Metal as well as the continuity or "Handoff" to OS X and doubling the RAM in Air 2 TRULY shows off the capabilities of the second gen A-series processor/SoC with IT's graphic set --- then. 2015 Apple's third revision of the 64bit A9 processor sporting different IT GPU options across devices (iPhone --> iPad Pro) and it's near doubling in both compute and graphics, doubling of RAM across the board and the real and perceived speed differences on this generation's iOS devices compared to last is truly incomparable!
    I owned the 6+, I upgraded to the 6s+. The difference is extreme.
    I upgraded a year ago from Air 1 to 2. The differences are incredible --- including the current gotta have features of the new devices with slide over, dual app and video overlay without latency or delay is phenomenal and a feature set that may work in some capacity on the older devices but limitations are in place so as joy to kill the device.
    One thing Apple maybe too good at, keeping the older devices working and with most updates the latest OS allows. I've got a ½ dozen iPad 4s we use in the field and not a single one has needed to be fixed --- & they've been through hell and back. I'm replacing them with the Air2s x4 and a pair of iPPs as the ONE thing the iPad does have over the rMBP's - constant connectivity. Always online and lighter, more durable with longer life than the rMBP (my favorite computer BTW, I bought the original 15" 2012 model, still going as strong today as the day I bought it. Wife needed an update to her laptop and just got her the 2.8/16/1TB CTO & I'm jealous! Amazing the speed difference between the older SSD ('12) models vs. the PCIe SSD solutions of today, I couldn't believe how quickly the Adobe CC suite downloaded (we're blessed with local gigabit ISP service) and then installed. Transferring larger files is mesmerizing. Using Thunderbolt or usb 3, it's nearly instantaneous!
    The new rMBP (not so new anymore, but they're definitely building out the HiDPI displays and knocking them outta the park) is a revelation in my 30 years of computing. Starting on the IIe and transitioning to s 286 in corporate world, back to Mac in 2006 for my business haven't looked back. We keep a Windows machine for a Quickbooks machine and a few older games my son enjoys as well as the Xbox integration ...but OS X is certainly a phenomenal operating system in conjunction with iOS, it's mobile counterpart along with Handoff creates the perfect vertical and horizontally integrated and aggregated system we've ever had, without the need for a four year computer programming and networking degree --- Apple has designed a total system of work and play that simply keeps everything. Without the need to even 'think' about what you're doing
    While still retaining the power of Terminal, Automator and Xcode free for all to simulate or build their own iOS or OS X apps themselves!
    I believe MS is poised to challenge Apple and dethrone Samsung, as well as Android (no one goes home to their Chromebook to finalize the rendering of their motion captured that day, manipulate their photos in batch or finish their 3D composites they started with Adobe on their tablet earlier in the day.
    Windows 10 is a decent OS and its ability to integrate and aggregate with its mobile platforms will eventually topple Android as the #2 mobile option challenging OS X for the top spot.
    Time will tell but Apple has a tough time releasing shitty products these days. I wasn't sure about the Apple .watch my wife got me for my birthday. Today I can't imagine life without it as there's a couple dozen times a day my phone can stay in pocket, on desk or away from my immediate access while the watch tells me what I need to know.

    Back to Macrumors, the debate(s) as there's several are spirited and with merit. Plenty of very informational material and links with the back and forth, very interesting debate to be sure as who'd have thought the debate itself would be happening this early? Apple has money. Apple poached plenty of chip experts many years ago, as well as low level programmers and SoC designers and fabricators --- it's been quick but Intel has rested on its laurels a long time with AMD's woes and complimentary discrete GPU OEMs. It's only been since the X86 architecture replaced PPC on OS X seemingly that Intel has developed an interest in building up the integration of GPU on their CPU builds.
    And only the last couple years, since the Intel 3000 has IrisPro made the leap that DOES challenge Imagination Tech's formidable and mobile graphic solutions

    ...again, JUST to be having this debate says plenty about the job Apple is doing with their silicon
  • zeeBomb - Wednesday, November 11, 2015 - link

    Well this came earlier than I expected. My moneys on they're using a 10 cluster gpu as well, or the GT900.
  • dnzk - Thursday, November 12, 2015 - link

    My money on they're using customize GPU 12 cluster 2xGT7600, like on A8X with 8 cluster 2xGX6450.
  • V900 - Wednesday, November 11, 2015 - link

    "With the iPad Air 2, for example, I felt quite strongly that while in some ways the larger display was useful, it was difficult for me to justify the cost of the tablet due to the lack of some productivity-focused tools".

    This didn't make sense to me... The iPad Air2 has a bigger display?!? This is certainly news to me.

    As is the lack of productivity tools. Not only is the full Office out on iOS, you can even get Autocad these days...

Log in

Don't have an account? Sign up now