A73 vs A72

[Nintonito]

From early Kirin benchmarks it looks like overall performance is not much faster. Instruction decide width is down to 2 from 3, but apparently the rest of the pipeline is faster. And critically, ARM claims this core is designed to sustain itself at higher frequencies for extended periods of time.

[Helios]

Inside source tells me it's one more 'A' faster

[Helios]

In all seriousness, I'll believe it when somebody has a device in their hands and runs long term perf benchmarks.

This is always said literally every time the time comes for a new SoC, and it's always disappointing

[zxcvbad]

Not impressed by thev Kirin benchmarks, ARM design looks like a minor improvement. We'll see how it stands in sustained performance though in GB4 single-core bench it gets about 2K

[degasus]

AFAIK ARM called the A73 slower compared to A72, but more power efficient. So if you're bottlenecked by power (as common), you should get higher clock speeds with the A73 and so also a better performance.

[Grey1213]

(10-24-2016, 05:45 AM)Helios Wrote: In all seriousness, I'll believe it when somebody has a device in their hands and runs long term perf benchmarks.

This is always said literally every time the time comes for a new SoC, and it's always disappointing

Nah this seems to be a proper focus on it, the a72 wasn't even designed with mobile in mind iirc. But this is a change in focus that would take more than one year to implement. Personally I do expect it to get better but how much it gets better may not be that big a deal. Over the next few years it should be something substantial though... Hopefully.