I have to admit that I found I hard to listen with that camera going around in the silicon lab. Do we know anything more today that we didn’t know yesterday on how A14 compare to A12x?
That 11.8 billion figure will definitely increase with a14z and mac SOC. Also the a13 already scores 6500 in Geekbench compute - I doubt the a14 would be slower.A14 11.8 billion transistors, 6 CPU cores, 4 GPU cores, 40% faster CPU than A12, 30% faster GPU than A12, Geekbench 5 single core 1557, multi-core 4012, Metal 6033
A14Z 11.8 billion transistors, 8 CPU cores, 8 GPU cores, 40% faster CPU than A12Z, 30% faster GPU than A12Z, Geekbench 5 single core 1565, multi-core 6483, Metal 13438
A14Z Mac SoC 11.8 billion transistors, 12 CPU cores, 32 GPU cores, 150% faster multi-core than A14Z, 400% faster GPU than A14Z, Geekbench 5 single core 1565, multi-core 9725, Metal 53752
That would put A14Z Mac on par with 10-core i9 in iMac 2020 and the GPU between RX 580 and RX Vega 56.
I have to admit that I found I hard to listen with that camera going around in the silicon lab. Do we know anything more today that we didn’t know yesterday on how A14 compare to A12x?
...unscientific speculation...
A14 11.8 billion transistors, 6 CPU cores, 4 GPU cores, 40% faster CPU than A12, 30% faster GPU than A12, Geekbench 5 single core 1557, multi-core 4012, Metal 6033
A14Z 11.8 billion transistors, 8 CPU cores, 8 GPU cores, 40% faster CPU than A12Z, 30% faster GPU than A12Z, Geekbench 5 single core 1565, multi-core 6483, Metal 13438
A14Z Mac SoC 11.8 billion transistors, 12 CPU cores, 32 GPU cores, 150% faster multi-core than A14Z, 400% faster GPU than A14Z, Geekbench 5 single core 1565, multi-core 9725, Metal 53752
That would put A14Z Mac on par with 10-core i9 in iMac 2020 and the GPU between RX 580 and RX Vega 56.
Why is everyone ignoring this?:
"... Building upon this architecture, Apple is designing a family of SoCs for the Mac."
(Source: https://www.apple.com/newsroom/2020/06/apple-announces-mac-transition-to-apple-silicon/)
For me, it clearly means we will get custom chips for Macs. ?♂️Because it doesn't mean anything. It could be a new chip, it could be the same chip as the iPhones/iPads, or it could be combination of both. I can certainly see Apple using A14X in a 13" MBP and something much more powerful in a 16" MBP for example...
A14 is important because it will set the expectations for the Mac hardware. If the A14 does not contain any new architectural features, it would be unreasonable to expect them on the Mac side. And it gives us a good estimate on how well new Macs will perform.
Interestingly enough, the little information Apple gave us is consistent with the earlier leaked A14 Geekbench scores. Still, we will have to wait two weeks until more in-depth analysis can be done.
For me, it clearly means we will get custom chips for Macs. ?♂️
Don't know. But not to long and we'll see. ?What does "custom" mean? Is it the same custom as A12X vs. A12 (same exact architecture but more cores and cache)? Or does "custom" imply deeper architectural changes here?
I think that from Apple perspective, reusing as much stuff as possible would make the most sense. My suspicion is that "custom" simply means same building blocks, but more of them. The iPhone and iPad Air use 2 high-performant A14 cores, a Mac chip might use four or eight. I seriously doubt that they would develop a new sub architecture just for the Mac.
So no big leap compare to A13 but maybe less power consumption?
At this point Anandtech does not have more information than us. We don't know whether A14 cores are any different, what clocks they run on and what kind of power consumption they have.
Another, very important point: computer chips are usually performance binned. Due to fluctuations during the manufacturing process, otherwise identical chis will inevitably end up having different characteristics. Some can maintain higher clocks, some will consume less power at the same clock etc. etc. Binning refers to the process of testing the produced chips and sorting out which one's are "better". The few best-performing chips (e.g. able to sustain higher speeds) are then sold at a premium. Mobile chips instead are power binned — they are checked for how much energy they required and then fine-tuned (in software) to only go until a certain limit. So if you get lucky, you might get a chip that performs 10-20% or so faster at the same power draw, not really noticeable in a phone or a tablet. Power binning usually implies that the speeds are conservative, since you don't want to throw away chips that work fine just are a bit sower.
For Macs, it would make a lot of sense to use performance binning. Chips capable of better speeds (potentially at a sight expense of power consumption) will be a better fit for a Mac laptop, while chips that use less power could go into an iPad Pro. This is also a great way for Apple to save money.
This is why I expect first ARM Macs to run on the same chip as the upcoming iPad Pro, but using higher clocks.
The A14 and all of the Apple Silicon processors that preceded it are systems on chips. The iPad Pro and Macs are completely different systems. So, no....
This is why I expect first ARM Macs to run on the same chip as the upcoming iPad Pro, but using higher clocks.
Well, I'm no expert hence "the unscientific speculation", but A12Z has also more CPUs and GPUs than A12 but the same number of transistors so that's why I wrote the same number. Also there is no other information about the transistor count of the Mac Soc so I didn't know what else to write.If you are adding CPU & GPU cores, but keeping the same 11.8 billion transistor count, what are you removing from the rest of the chip...?
Well, I'm no expert hence "the unscientific speculation", but A12Z has also more CPUs and GPUs than A12 but the same number of transistors so that's why I wrote the same number. Also there is no other information about the transistor count of the Mac Soc so I didn't know what else to write.
That 11.8 billion figure will definitely increase with a14z and mac SOC. Also the a13 already scores 6500 in Geekbench compute - I doubt the a14 would be slower.
I would also imagine with larger thermal envelope and active cooling - the Mac SOC cpu would have clock higher speeds - scores should increase.
Strange! The Wiki page I used as source said 6.9 billion. In that case A14Z should have 17.1 billion transistors and Mac SoC many more.According to Apple, A12 has 6.9b transistors while A12X/A12Z (its the same chip) has around 10b transistors.
pretty sure you mean 50 and 300 respectively150% faster multi-core than A14Z, 400% faster GPU than A14Z