hattig

LanceMeAlot said:

I have a mackbook air M1 for working purposes. For office, and online work it’s fantastic but as soon as you use MS teams for a meeting, sharing a screen and video conferencing, the cpu get’s too hot. And because there is no fan inside MacOS makes the cpu’s work slower to avoid heat problems. And then the machine will get very slow. Therefore I would always buy the 13” mackbook pro to avoid this problem.

MS Teams does that on 16 core behemoth desktop systems - it's not the system, it's the appalling software from MS.

If you make your own silicon designs, then you have two costs - upfront research, design, testing, and then manufacturing.

The more you sell, the more you amortise the first cost over all of the sold devices. Apple is already investing the money into the first part for the iPhone/iPad SoCs - it is an incremental additional cost to do a new design for Macs. But let's say it's $500m for a 5nm design at this stage. Apple will ship 16m Macs a year. That's $31 per Mac. It'll be lower because a lot of that cost is shared with iPhone SoC. 

A 5nm wafer processing cost might be $12k (dropping as it matures) right now. It's 80% denser than 7nm. Let's assume they use that for the additional CPU cores and GPU cores the desktop chip would have and the resulting chip is 10.1mm x 12.6mm just like the A12Z. Even with poor yields you get over 300 fully working die per wafer out of ~450. That's a fabrication cost of $40 for each fully working die (and you'll get partially working dies as a bonus as well, for your lower-end products).

I imagine that Apple want to grow the Mac business and will use lower priced Macs with those recovered dies to achieve that - although there's a certain level of quality that Apple won't go below.

How much does Apple pay Intel for the chips in their laptops currently? Sure, it's discounted over list price, but it will be far, far higher than $71.

These numbers are extremely rough estimates.

Large amounts of RAM are needed for slomo functionalities. The (Android) 48MP+ Camera sensors embed 512MB (maybe more in more recent designs) of RAM into the sensor to achieve this, as the data link to the SoC is not fast enough. However the iPhone may not have this limitation, so it can use system RAM for this functionality, as well as standard camera functionality. The Camera app could also be kept in this RAM permanently, for even faster instant-on. Also the AI cores used so often in photography will have access to this RAM for their functionalities, making system RAM desirable for this use.

The memory is most likely surface mounted LPDDR4X. The GPU would suffer otherwise.

I expect the A14Z (or whatever they call the Mac Apple Silicon) to utilise LPDDR5 and not be upgradeable in the slimmer form factors.

Early devices might use LPDDR4X still. These memory controllers usually get slower DDR support as well, so DDR4 for the expandable devices (DDR5 when that becomes available - later than LPDDR5). I don't think HBM will be used in the first generation, but if Apple ever go large on the integrated GPU this may be their only option when that happens.

It'll almost certainly integrate USB4 for Thunderbolt functionality. If not, they'll dedicate some PCIe lanes for an external Intel thunderbolt controller. It's possible Apple will split the chip into an I/O die and the compute die, but on the same package.

PCIe will have to be included, it'll have to be PCIe4, and I would expect 4x for SSD and 8x/16x for discrete GPU (for devices that need more than what the Apple GPU can do internally), and maybe another 4x for other devices/storage.

There's a good chance that it'll be an ARMv9 ISA, but not 100% - but certainly by the time the Mac Pro migrates.

TBH the main thing remaining is for Apple to integrate the modem into the SoC. Samsung only use Snapdragon in the US because of the modem differences there. Exynos is everywhere else. Exynos really is the closest chip in spirit to Apple's (because only Samsung use them these days), but Samsung don't appear to care too much about leading the market, instead being content to roughly match the equivalent Snapdragon in performance. The M1, M2 cores that Samsung uses appear to be lightly modified basic ARM core designs. Even Qualcomm have gone this direction with their latest Kryo cores. Apple have taken control of their chip design future, and have the money and the sales to drive this forward for the foreseeable future. It's unlikely that ARM is going to launch a mobile core that will compete on single-threaded performance anytime soon, and now that Apple have fixed their multi-core processing to use all cores, even the traditional "octo-core" method for other SoCs to get great scores has been surpassed. I can't see any other SoC outperforming the A series SoCs in CPU/GPU for many years (Kirin 970 on-paper has a more powerful AI core, we should find out more within a month or so, I note an AI core is only as good as the learning applied to it which Apple surely has not skimped on). Still, performance is not everything in a phone, but it's not like A11 appears to have compromises in other places like battery life. Like many, I would really like to see an A11 or later SoC (likely a different enhanced design that is optimised for this usage) running OS X on a laptop. Let's see what the additional TDP headroom can do in these form factors.

"fell behind Samsung's Galaxy Note 9 (powered by an 845)" Not with that score, that must be an Exynos version with its custom Samsung Mongoose ARM core.

What an absolute bellend.

blastdoor said:

If Geekbench is running through Rosetta then these numbers are actually pretty impressive. 800/1100 = 72% of full performance when running emulated code? That's really impressive. 

That's right.

Also consider that Geekbench is a benchmark - it's a pure x86-64 kernel running in isolation, so it's 100% worst case scenario.

Most applications will call out to the system to do even the simplest of things - which is pure native.

Note in the State of the Platform show, they do confirm it's ARM in the Rosetta 2 section, for those doubters out there (or the RISC-V theorists!).