anonconformist

chadbag said:

Why would a unified memory system be a detriment to memory intensive operations?

GPU cores needs to contend with CPU cores for memory cycles. Unless there’s fast enough memory to fulfill all CPU requests and GPU requests at the same time, something will need to get memory access at a lower priority. If the CPU cores have a lower priority, the CPU cores are stalled and doing no computation while waiting for the GPU cores.

Note also that main memory requests haven’t been for a single byte or even 8 bytes at a time for a very long time: all main memory access is done at chunks the size at least of CPU cache lines. I don’t know what the M1 uses, but for at least the last 20 years the smallest cache line has been 32 bytes.  It’s done this way because the electrical signaling overhead takes a meaningful amount of time, and RAM (dynamic) also has other costs, but the net result is a sequential chunk is usually going to be far more efficient to access than 8 bytes or less.  The latency between main memory access and getting/setting data is many many CPU cycles as main memory is a fraction of the speed of CPU L1 data and instruction caches, which tend to be slower than CPU register accesses.

By combining GPU cores, CPU cores and huge L1/L2 caches along with L3, and minimizing protocol overhead between CPU and GPU along with minimal wire time, at least (assuming they’re not using buses between them like PCIE) they can make this negotiation between the various SoC function areas very fast, far faster than if it were spread out on a motherboard.  Despite all that, it doesn’t change the fact that it’s unlikely that all parts that need memory access can have as much memory access at the same time as it’d take to keep them all running at maximum performance.

tzeshan said:

Apple will be able to eat Windows OS shares this time. Two things. Apple Silicon Macs are rumored to last up to twenty hours on battery. And Pages, Numbers, and Keynote free applications are mature enough to be able to replace Word, Excel, and PowerPoints applications from Microsoft which are very expensive to own. 

1. Incredibly few people need 20 hours of battery life for a computer, because very few people are working away from a power source that long. 20 hours battery life is a nice thing, in theory, but in practice, something else has to give to get it. Note Apple has kept iPads at about 10 hours battery life from the start, with * as to “10 hours” where most people that need heavier computation needs are likely plugged into an outlet anyway.
2. For home use, Apple’s office suite is sufficient: with what’s expected at a lot of businesses, there are huge Numbers of spreadsheet functions used daily in Excel not provided by Apple’s software, and Pages of word processor functionality in MS Office is the same way.
3. All that being said, Silicon Macs will inherit a huge library of iOS/iPadOS software, even if suboptimal to use on other than a touchscreen, and for very stupidly-cheap prices for functionality, in fast hardware. If Apple announces touchscreen Macs (they’ve been known to say “we won’t do this! Then go and add it: Apple Pencil, anyone?) then things get really interesting,
4. SwiftUI (at 2.0 with a lot more functionality, but still needs work) makes it far easier to have cross-Apple-platform applications with amazingly little added work that look/feel native, because they are in the GUI frameworks they’re using, to a large extent. With Apple Silicon Macs testing is capable of being done on the same hardware quirks to enough certainty that it greatly eases the burden of supporting multiple types of Apple devices, and increases the potential sales opportunities.
5. An issue regarding platform viability for making a living via Apple platform apps are customers believing all software needs to be cheap or free, and how DARE developers try to make a profit!

tht said:

mjtomlin said:

It's rumored to sport the A14X processor, a custom GPU, and a battery life between 15 to 20 hours.

Umm, who started that rumor? And why am I hearing that the first ASi Mac will have an A14X in it? Apple has already said they were working on a new family of SoCs for the Mac.

I don't think it was Gurman or Kuo. Both of them are pretty careful about that level of specificity. It was probably your Twitter leaker du jour. The names don't matter as it's an artificial distinction anyways.

A notional A14X with a 4+4+8 config: 4 perf cores, 4 eff cores and 8 graphics cores, would be awesome as one of the processor options for an MBA, MBP13, Mac mini, small iMac, and whatever low cost machine as well as an iPad Pro. It just needs to support up to 32 GB of RAM and TB3/USB4 for the Mac machines.

An A14 based MBA13 with 8 GB of RAM and 128 GB of storage for $800 would make for a great machine too.

But yes, the higher performance SoC options will be "made for Mac" as it were. They will run too hot to be in a fanless form factor.

These days for a MacBook Pro or Mac desktop supporting up to 32 GB is insufficient. I bought my 2019 16” MBP with 64 GB and I have no reason I’d go backwards to less (yes, I can absolutely use all that, though currently I won’t peak on a daily basis).

wizard69 said:

rob53 said:

verne arase said:

Good.

Let them make their own cake.

Totally agree. Why should Apple support other computer companies? All this would do is give Congress another reason to investigate their monopolistic activities. 

Actually congress could investigate them for not selling the processors.  Congress can investigate Apple for any imagined or real excuse they can come up with so Congress has nothing to do with this issue.

The best reason for Apple to start selling chips is to have a wider base to spread development expenses across.   Apple could literally get silicon customers to pay for the development process.

Your argument makes no business sense with Apple’s objectives: they could make money on the chips, but then the other competitor devices would be able to undercut them on cost on the total device and thus cost them profits on Apple devices, while also anchoring them to producing contracted amounts of chips for their competitors as a legal priority before fulfilling their own device needs, lest they befall the extreme danger of becoming a regulated monopoly.  The reasons Intel wants and needs AMD to survive and do well enough to be competitive at minimum are:

1. Many entities need to have a backup option for replacement hardware, even with a little hardware design change, that performs comparable to what they already have, should their main supplier have a problem.  Many government contracts have this.  Apple in some unknown number of parts has more than a single supplier for supplier diversification for this reason, even if their regular supplier can (normally) provide for all their needed parts.
2. If only a single Intel X86/AMD64 ISA chip provider existed, they’d become a monopoly and subject to all those regulatory and other encumbrances that result.
3. AMD hasn’t been truly competitive more years than they have been mediocre: Intel has become complacent as a result. Real technical competition is needed in the market to keep technology progressing.  If AMD didn’t exist, I believe other ISAs (more than ARM exists) would have eaten Intel’s lunch sooner, because of the Intel tax.

ne1 said:

“...unleashed another round of innovation.” 

This is a VERY interesting sentence— he seems to imply that there are exciting cross platform macOS / iOS  products or software on the horizon. 

I’m trying to figure out what future product/capability he’s implicating. 

By going to Apple Silicon (and really, all their other smaller devices already are using low-power versions) it’s trivial to have the only significant differences in capability for devices being speed and user interface: the guts are almost identical to those machines, with only minor device driver differences and hardware differences for things like touchscreens.  When you can count on all the device SOCs having the same AI hardware support, crypto support, whatever it is, it allows a far more unified OS/device capability and makes it even easier to lure developers to develop for Apple devices in general: to a fairly large degree (but not completely) with SwiftUI and the hardware commonalities it means developers can reach far more potential customers for a given amount of learning the platform.