Qualcomm's Snapdragon 810 SoC has not had the easiest of introductions. Some handset manufacturers have been reluctant to commit to it and depressed Qualcomm's order book. Others looking to use the chipset have delayed production and set release dates later in the year. Major partner Samsung switched the entire Galaxy S6 family to its home-grown Exynos system - previously Exynos would power South Korean handsets while Snapdragons would be in the handsets shipped around the world.
What are the actual issues? With the HTC One M9 and LG G Flex 2 on the market, it's possible to test the Snapdragon 810 system to find out how the new chips handle excess heat, CPU throttling, and how it compares to other options. Ars Technica's Andrew Cunningham reports the details, and his summary is clear:
In short, chips throttle, but the 810 throttles more than most, and it's severe enough that the 810 is actually slower than the 801 or 805 in some CPU-bound tasks over the long haul. The Exynos 7 Octa, which has similar specs on paper, is much better in practice.
At this point, Qualcomm has implied to us several times that its use of ARM Cortex CPU cores was a stopgap measure...
Why has Qualcomm dropped the ball and rushed the weak Snapdragon 810 to market? Why are manufacturers having to choose between clocked performance, delays to release dates, or having to look for solutions outside of the until-now-reliable Qualcomm?
Because of Apple.
Tim Cook's announcement of the A7 chip in the iPhone 5S is the key moment in this narrative. By pushing iOS to 64-bit at the 'wrong' time in Qualcomm's product cycle, Apple created a fork of grandmaster proportions for the competition to deal with. Qualcomm's new-design 64-bit chip (now labelled the Snapdragon 820) will not be generally available in handsets until the first half of 2016, so the manufacturers had to choose; stick with their roadmaps and release a 64-bit handset years down the line (handling a marketing advantage to Apple), or to do what Android does best and play 'Specifications Trumps' by matching the 64-bit marketing machine with solutions that had a hint of the MacGyver about them (handing a technical advantage to Apple).
I've spoken before about how the iPhone 5S's jump to 64-bit computing pushed Android down the 64-bit software route faster than planned. Now these benchmarking numbers show that the hardware has had to face the same issues as the software:
What was left unsaid was the corporate advantage that the jump to 64-bit offered Apple. At the time there was no comparable option for Android smartphones to jump to a 64-bit processor – and even if there was hardware on the shelf capable of bodging a 64-bit architecture, Android OS itself would not offer 64-bit out of the box until Android 5.0 Lollipop was released.
With Android software pushed ahead to 64-bit, the hardware had to follow in short order. Qualcomm was bounced into offering a 'quick' 64-bit Snapdragon alongside the roadmap it already had. The answer was the Snapdragon 810, and as you can see, it has all the marketing benefits of 64-bit, but fewer of the technical benefits.
Apple's A7 was faster than its predecessors in CPU-based tasks, faster in graphical tasks, and offered longer battery performance. All of these benefits applied to the iPhone 5S and the engineering knowledge will have been passed on to improve the A8, A8X, and presumably the upcoming A9. Apple could continue developing the iOS software and mobile hardware at its own fast pace.
Which is exactly the opposite case for Android. In the rush to match Apple in the 64-bit stakes, all of the advanced Android handsets have had to accommodate greater than expected compromises. Not only did Apple's move increase the potency of the iPhone, it bounced Google into rushing the 64-bit version of Android, it pushed Qualcomm to disrupt its product roadmap, and it weakened the Android handsets going on sale during 2015.
As Apple punched forward, it left Android with a headache that will impact on Google's platform well into next year.
