I find it weird that they're critizing the CPU for being old, as if age itself is a bad thing.
I find it hard to believe an OMAP3 at 1GHz would be insufficient to power a smartwatch. Perhaps something is wrong with the software, they're using slower memory chips, or the NAND is of the slow & cheap variety.
I currently have a 1GHz MIPS-powered Android smartwatch (pre-Wear though), and it's plenty fast. Battery life is not great (1-2 days), but this is a cheap SoC, produced on 45nm, with WiFi and bluetooth running all the time.
I'm honestly quite amazed at how much technology we manage to push into such small, battery powered packages!
> I find it hard to believe an OMAP3 at 1GHz would be insufficient to power a smartwatch.
That's not the claim being made, especially given it's compared to similarly-clocked SoC.
The article makes two claims: 1. the watch stutters which may be caused by the low floating-point performances (benched very low), and 2. the SoC is power-hungry leading to miserable battery life.
I find that floating point story hard to swallow. It looks like they went "huh, this CPU benchmarks roughly as fast as the others, but we find it slow in practice. Only floating point benchmarks lag behind the competition, so that must be why!".
Floating point instruction are rarely used in common software. They're important in rendering (which is offloaded to a GPU in this case) or scientific computing (which is irrelevant here, unless you were planning to fold genomes on your watch).
It's possible that this OMAP came without an FPU to save space and power.
Most Linux software for ARM is compiled with softfloat anyway, because it rarely matters for end users, and lots of ARM boards don't have an FPU. Some distributions have a separate armhf version, which does use hardware floating point operations.
The quoted /proc/cpuinfo reports vfpv3 and neon, so it does have an FPU (and it even has the Neon SIMD instructions, which not all SoCs of that era do).
From my perspective the softfloat-vs-hardfloat balance has tipped and I would say that "most Linux software for ARM" is now hardfloat rather than softfloat. But your view on this kind of thing depends a lot on whereabouts in the embedded world you are: if you're at the trailing edge then you'll see a lot more ARMv5 non-FPU devices.
Exactly the case, and it really undermines the whole article. The general floating point performance (non GPU) is entirely irrelevant for the general uses of this device. Other subsystems might be faulty, or it could be a faulty device, but it certainly isn't floating point performance.
One line in the article that gave me a chuckle was We can say that Android Wear looks much better on a round screen. The OS was clearly designed with circular displays in mind—every button is round, producing a cool circle-within-a-circle look on the 360. Contrast this with the Verge who said that Android Wear was designed and looks better on a square screen, and is forced onto a round screen.
>I find it weird that they're critizing the CPU for being old, as if age itself is a bad thing.
When you are cramming an old CPU into a tiny, cutting edge device like this, I think the criticisms might be warranted. Especially so considering the poor battery performance of the watch. It might not be entirely due to the CPU but it seems probable that the CPU is a nonnegligible contributor to the power consumption issues.
I find it hard to believe an OMAP3 at 1GHz would be insufficient to power a smartwatch. Perhaps something is wrong with the software, they're using slower memory chips, or the NAND is of the slow & cheap variety.
I currently have a 1GHz MIPS-powered Android smartwatch (pre-Wear though), and it's plenty fast. Battery life is not great (1-2 days), but this is a cheap SoC, produced on 45nm, with WiFi and bluetooth running all the time.
I'm honestly quite amazed at how much technology we manage to push into such small, battery powered packages!