blazorthon
Distinguished
[citation][nom]obsama1[/nom]Apple can't keep putting 800MHz processors in their phones. Samsung is coming out with a 2GHz Exynos dual-core soon! Come summer, and we'll have 1.5GHz-2GHz quad-cores in phones! Also, I agree with calling it the iPhone 6.[/citation]
Clock speed is only one factor in performance. Linearly increasing clock frequency (say, doubling it from 1GHz to 2GHz) will increase power usage in a more exponential manner (it will use more than twice as much power despite having only about twice as much performance). A higher clock speed is only relative to higher performance for the same CPU.
For example, a P4 at 3GHz will have a near 33% jump in performance going from 3GHz to 4GHz (an ~33% jump in clock frequency), but all of that clock speed isn't enough for it to not be hammered by a 1.6GHz Sandy Bridge Celeron.
How does this apply to mobile CPUs? Well, in the exact same way. A Cortex A15 dual core at 2GHz will hammer a quad core 1.5GHz Coretex A9 because the A15 has vastly superior IPC (Instructions per clock, aka performance per Hz) due to it being a superior architecture.
A dual core Cortex A15 at 800MHz will probably not be better than a quad core Tegra 3 (Cortex A9) at 1.5GHz or so, but it will be better than a dual core Cortex A9 1.5GHz dual core CPU. TI has a CPU that is about 3 times faster than Cortex A9 at the same clock frequency, so if the A5X is anything like that, it will be about 75% of the Tegra 3's full CPU performance, except as a dual core with a much lower frequency (and thus power usage and cost) and thus is more easily loaded up with Android apps because it has fewer threads to worry about to keep performance at maximum efficiency.
Moving on from that, we might be seeing Androids with monstrously fast processors, but this is an iPhone here, not an Android. Most users of either camp won't utilize such processing power on these devices, especially on the iPhone side, so it would not be surprising to me nor even really a bad thing to see the next iPhone an under-performer compared to the Androids. I would never buy any of the iPhones regardless unless I had to for some reason, but please consider the context here.
Clock speed is only one factor in performance. Linearly increasing clock frequency (say, doubling it from 1GHz to 2GHz) will increase power usage in a more exponential manner (it will use more than twice as much power despite having only about twice as much performance). A higher clock speed is only relative to higher performance for the same CPU.
For example, a P4 at 3GHz will have a near 33% jump in performance going from 3GHz to 4GHz (an ~33% jump in clock frequency), but all of that clock speed isn't enough for it to not be hammered by a 1.6GHz Sandy Bridge Celeron.
How does this apply to mobile CPUs? Well, in the exact same way. A Cortex A15 dual core at 2GHz will hammer a quad core 1.5GHz Coretex A9 because the A15 has vastly superior IPC (Instructions per clock, aka performance per Hz) due to it being a superior architecture.
A dual core Cortex A15 at 800MHz will probably not be better than a quad core Tegra 3 (Cortex A9) at 1.5GHz or so, but it will be better than a dual core Cortex A9 1.5GHz dual core CPU. TI has a CPU that is about 3 times faster than Cortex A9 at the same clock frequency, so if the A5X is anything like that, it will be about 75% of the Tegra 3's full CPU performance, except as a dual core with a much lower frequency (and thus power usage and cost) and thus is more easily loaded up with Android apps because it has fewer threads to worry about to keep performance at maximum efficiency.
Moving on from that, we might be seeing Androids with monstrously fast processors, but this is an iPhone here, not an Android. Most users of either camp won't utilize such processing power on these devices, especially on the iPhone side, so it would not be surprising to me nor even really a bad thing to see the next iPhone an under-performer compared to the Androids. I would never buy any of the iPhones regardless unless I had to for some reason, but please consider the context here.