Apple dumps Intel and shows Final Cut Pro on new Mac with Apple CPU !

Provided it's all true, that is a product of dedicating your life to your career or field of expertise. [Every industry has them.]

What separates the knowledge from a basic vision of building computers (which isn't difficult) is not only the deep comprehension for the hardware, but also the understanding of the technology's history and possible future.

We can speak about cameras in a similar fashion because we know what has been done, which new specifications might be unprecedented [like 4K/240p full-frame medium-format RAW with DPAF, IBIS and electronic ND...although that doesn't sound that impressive anymore, LOL], and predict what camera creators may, may not, and could do while also understanding their competition and the overall entire state of the industry.
 
P.S. Without knowing exactly what type of products Apple has planned, it's impossible to really make any conclusion from the analysis. The main takeaway at the present time is their developments are impressive (even with any potential hurdles), but a very strong competitor - on the rise along with you - provides a blueprint that is subject to change.

The write-up sounded how I always pictured the contested computing industry.
 
After reading through this thread I dug out my 2013 13 inch MBP that had been working fine until my 5 year old tipped a glass of water over it last spring. Booted up with no issues. Currently updating. Gotta hand it to Apple - their machines definitely have a long life!
 
NorBro said:
P.S. Without knowing exactly what type of products Apple has planned, it's impossible to really make any conclusion from the analysis...
Click to expand...
The poster compares Apple's iPhone technology with the Intel technology. A reasonable analysis from the hardware engineering POV.

Howevah - he doesn't offer any corporate insight (kudos to him/her for avoiding something he doesn't know). I'd say, with the utmost certainty, that Apple wouldn't have dumped Intel without completing the necessary steps on the prescribed path. The "maybe" is not good enough at that level. For Cook (and the board) to have made this decision, they had to be sure about where they stand with the development of their own technology. Furthermore, one has to assume that, since the Arm based processors are good enough for the supercomputers and the massive server clusters, that that's where Apple is headed. Mac, iPad is a minuscule portion of the business they're aiming to conquer.

The mobile phone industry is pretty close to hitting a limit - screens, processors, cameras - are maxing out on their utility to an average user. Apple sees the obvious and is making a big move into everywhere a computer is needed.

Which is everywhere indeed.
 
In the one statement he made it sounded like he was pretty confident about Intel never being able to produce something as good so I'm sure Apple feels good about their decision too, lol.

I'd like to read someone write more about AMD's future and how they may affect Apple's.
 
DLD said:
The mobile phone industry is pretty close to hitting a limit - screens, processors, cameras - are maxing out on their utility to an average user. Apple sees the obvious and is making a big move into everywhere a computer is needed.
Click to expand...

I'm not sure this is entirely true about mobile chips maxing out their utility. As we move into 5G, Augmented Reality... there may be new apps and functions that take advantage of more power in a way the average user finds useful.

I mean, I remember 20 or so years ago people saying, "once they come out with these massive new 64GB hard drives, you'll never need to buy another drive because you can fit everything you've got on there."

We have a similar debate over camera tech ("if you think 4K is great, just wait for 8K) but I don't think it's the same as computing power because there's a limit to what the viewers need. They can only ascertain so much resolution. They're only going to fit so large a screen and sit so close. Some uses would like unlimited resolution, but I'm still shooting half my stuff in HD. Once you have over a certain amount of dynamic range, the rest is wasted in 99.99% of shots. Some people would like a million fps but most of us never touch 120fps. We're not creating new uses because the physics of what we shoot and how we show it stays the same. Not true for video games or augmented reality. I mean even look at transmitting video - blu-ray is supposedly making a resurgence because people have figured out how heavily compressed their streams are
 
The mobile market is nearing the point the computer industry hit about a decade ago. For a casual user, the processor speed, the amount of RAM, HD/SSD, screen size/resolution have pretty much hit their marginal benefit points. (not to be confused with the gamers and the professional users) That's why product that used to average $1,500 is now down to about $500-$700 and the replacement cycle from 3-5 years to 10-12. There's, obviously, some room left with the SSD/Optane drives, GPU, etc. but those mean little to most. I mentioned this a while ago but there are a lot of folks who just own a smartphone and nothing more.

Besides, every product has its cycle. It's not known to everyone but the railroad industry that was the economic beacon for over 60 years of the 19th century experienced a lot of bankruptcies in the early 20th. It was overbuilt, over leveraged and over estimated. And that was before the car market began to flex its muscles with Model T. These days, very few travel by train in the US. The railroads are pretty much all freight.
 
Railroads are still an essential service in the Northeast. Besides NYC's subways, if you work in the city and live on Long Island or in the tri-state area of NJ/CT it's on your mind and part of your life almost everyday, but I guess I never thought about the rest of the country.

LIRR had 90 million passengers in 2018.
 
DLD said:
Let me opine on something I know nothing about ... oh, like usual - new ARM chips from Ampere, Marvell and Amazon could probably be tweaked - easily tweaked, perhaps? - to run on the similar architecture from Apple. In that case, Apple could co-produce some of them, rebadge or simply purchase whatever they need to fill out any gaps within their lineup.
Click to expand...

No... Apple doesn't actually produce any silicon, so that's not an option. TSMC is producing Apple's ARM processors.

PS. Since ARM chips are going inside pretty much everything these days, it moves Apple into a chip making business outside its own products and/or allows it to sell packaged hardware+software deals to whomever. Which means economy of scale and lower prices for everyone.
Click to expand...

No. That would involve Apple essentially resurrecting the clone program, which is the opposite of what it's looking to do here.
 
DLD said:
The mobile market is nearing the point the computer industry hit about a decade ago. For a casual user, the processor speed, the amount of RAM, HD/SSD, screen size/resolution have pretty much hit their marginal benefit points. (not to be confused with the gamers and the professional users) That's why product that used to average $1,500 is now down to about $500-$700 and the replacement cycle from 3-5 years to 10-12. There's, obviously, some room left with the SSD/Optane drives, GPU, etc. but those mean little to most. I mentioned this a while ago but there are a lot of folks who just own a smartphone and nothing more.

Besides, every product has its cycle. It's not known to everyone but the railroad industry that was the economic beacon for over 60 years of the 19th century experienced a lot of bankruptcies in the early 20th. It was overbuilt, over leveraged and over estimated. And that was before the car market began to flex its muscles with Model T. These days, very few travel by train in the US. The railroads are pretty much all freight.
Click to expand...

Watched a documentary on Penn Station and the associated railroads the other month and it was fascinating. Especially looking at how much and what all went into it and just how short of a time span it was used and relevant.
 
Tamerlin said:
No... Apple doesn't actually produce any silicon, so that's not an option. TSMC is producing Apple's ARM processors...
Click to expand...
Of course, who builds processors is irrelevant since they're built to order. They may not be built on volume required for other tasks but that also presumes than other foundries besides Taiwan Semi can't be contracted.

And the whole point is moot anyway, as Apple is going to build an own chip plant in Arizona. With TSMC. Construction's supposed to start in 2021 and production in 2024. And it requires a $12B investment. Clearly, Apple is going to need a lot more processors. I think I know why.
 
NorBro said:
Provided it's all true, that is a product of dedicating your life to your career or field of expertise. [Every industry has them.]
Click to expand...

It's pretty close. Some is optimistic; it's unlikely that any of Apple's current processors could approach the clock speeds of Intel's top end, partly because Intel's been tweaking the daylights out of its aging 14nm process and geriatric Sky Lake architecture for so long... heavily motivated by the fact that AMD's Zen architecture is simply more efficient, on top of having access to a more advanced process (which also helps efficiency).

There are a few obvious reasons that both Apple and Ampere use Intel as their point of comparison. One is that Intel's still the bigger dog, in that it has a LOT more money. Another is that AMD is quite a ahead. Intel's been pushing to reach 5GHz because that's the only way that it can top AMD's 4GHz processors. The last is that I don't think anyone wants to point to AMD right now, because of AMD's Infinity Fabric. That on its own is quite an accomplishment, engineering wise, and working with Cray is I'm sure a significant factor in why AMD's done such a good job on that side of things. Without it, the Rome would be either quite a clunker, or ridiculously hot for the perf.

Datacenter processors however don't necessarily need anywhere near the sophistication of Infinity Fabric; the vast majority of the workloads processors like Ampere are designed for aren't shared parallel applications, they're many independent processes, like web servers. They're rarely gated by compute, far more often by latency, since they'd be calling out to a database server for data access. Usually the relational DB servers are built on bigger machines with shared memory parallelism that you'd find on an engineering monster like the Fujitsu beast or an EPYC. And the power consumption on that Fuji monster is pretty big; there's a lot infrastructure in it.

The preliminary A14 benchmarks look like they're in reach of Sky Lake, but Intel's Comet Lake or Sunny Cove or whatever the <blert> the codename is for the new architecture is a HUGE jump in per-core and per-clock performance... maybe enough to put it neck and neck with Zen2, while of course AMD is gearing up to ship Zen3 on TSMC's EUV process... should be interesting. I don't doubt for a minute that AMD is aware of the process that ARM has made. I'm sure Intel is also, but Intel is also branching out because AMD's roadmap is designed to compete with Intel's pre-faceplant roadmap.

We can speak about cameras in a similar fashion because we know what has been done, which new specifications might be unprecedented [like 4K/240p full-frame medium-format RAW with DPAF, IBIS and electronic ND...although that doesn't sound that impressive anymore, LOL], and predict what camera creators may, may not, and could do while also understanding their competition and the overall entire state of the industry.
Click to expand...

And I keep trying to get out of that software world... but losing so much film work thanks to the pandemic drove me back :-/

Sigh.
 
DLD said:
Of course, who builds processors is irrelevant since they're built to order. They may not be built on volume required for other tasks but that also presumes than other foundries besides Taiwan Semi can't be contracted.
Click to expand...

That's not at all accurate. Apple can't produce processors for someone else, because it doesn't produce processors. TSMC can't produce Apple processors for anyone else either, because they're Apple's.

Apple's processors implement the ARM instruction set, but that doesn't mean much of anything for the rest of the processor. Apple's architecture beyond the decoding stage doesn't resemble Qualcomm's or Samsung's much, other than at a rather high level, and part of it is that Apple never cared about how much it cost to manufacture, because it isn't in the business of selling processors to integrators like Qualcomm and Samsung (technically so is nVidia, but there's just one customer for nVidia's ARM processors, and that's nVidia).

And the whole point is moot anyway, as Apple is going to build an own chip plant in Arizona. With TSMC. Construction's supposed to start in 2021 and production in 2024. And it requires a $12B investment. Clearly, Apple is going to need a lot more processors. I think I know why.
Click to expand...

Actually, there are some other dynamics also. One of them being that Apple is gaining a competitor for production capacity. In case you hadn't noticed, while Intel's been continuing to score record sales, AMD is STILL eating into its market share -- that tells us that the demand for processors is rising, so it's not just that AMD is taking market away from Intel (which it is). AMD is well on its way to becoming TSMC's second largest customer, and it's going to be manufacturing XBox and PlayStation cores in TSMC's fabs as well as PC processors and GPUs. Its older XBox + PlayStation parts predate the TSMC transition... I'm guessing AMD is still getting those made at Global Foundries, but the next generation, which are probably going to be going into heavy volume production soon if not already (Sony and MSFT want to sell them in time for the holidays, that means November, and that means having them ready to ship by October to get them into the distribution pipelines in time to get them onto shelves).

At the same time, nVidia's using TSMC to fab Ampere, and that's for its datacenter monstrosity that includes a 64-core Rome processor, lots of parts for auto manufacturers, and the suite of gaming and workstation products.

I'm very curious about how the new gaming consoles will affect the market. For years content creators have been riding on the wave of dirt cheap (for the perf + features) and insanely advanced graphics hardware driven by the gaming industry. Most gaming consoles are designed to be inexpensive rather than competitive with modern gaming systems at the time of launch. Microsoft and Sony are both going after high end gamers this time though, which means suddenly gaming consoles will be competing with high end gaming computers.

Mobile gaming will probably continue to be laptop oriented, but I have a feeling that nVidia's been anticipating this, and that the RTX Studio plus the push into computer vision applications (cars, autonomous drones, etc) and extremely high end datacenters is a response to a potential shift in the gaming industry that would rip the mid-range and enty level markets out from under nVidia entirely. nVidia doesn't have an APU option, and Intel's crappy GPUs are by a HUGE margin the most commonly sold GPUs on the market since they're built into every Intel processor, but Xe is... well, not crappy. (Faster than Vega, so far.)
 
DLD said:
PS. Some think that Intel is de facto finished or will be finished within a few years.
Click to expand...

No... that's very unlikely. Intel's not actually suffering financially. It's a huge company, so it can afford to branch out to a degree that AMD currently cannot.

What IS likely is that Intel will continue being second fiddle for quite a while on the CPU side of things, while it ramps up its GPU, memory, storage, FPGA, and modem production, among other things.

I think Intel will continue to make CPUs, but AMD 's squashed the margins on those. That said, those margins have been shrinking for some time, and Intel's been keeping them high because it could, not because it needed to. Intel already knew that GPUs were a better bet for the future than CPUs, hence the massive investment in visual computing and recruiting the Radeon Technologies Group's chief architect...
 
Intel just presented their first stacked CPU. That's a big deal, and, looking forward, a big advantage vs anybody else making CPUs nowadays.
Apple's iphone CPU is faster than what every other phone maker is using, see this and also scroll down here.
nvidia has an almost-monopoly on the GPU-compute world because their CUDA strategy was very successful and right now the most popular tools run on their GPUs but not on AMD's (e.g. tensorflow).
AMD has a very solid position right now on the PC market because they have a very good core design, and a very scalable packaging strategy.

I think they'll all be around for a while yet...

Regarding the original topic... I just bought myself a Threadripper 3970x workstation. Total cost for the tower was right under 4.000 eur. The mac pro that comes closest to my specs is 17.000 eur.
Apple is leaving the high-performance market anyway, and for the market in which they are strong, ARM processors are probably a good idea.
 
The articles I linked provide a different analysis of the chip market. (and, yes, it's OK to disagree ... that's what makes the world go around ... people make mistakes even when they don't ... or Quibi would be a runaway success)

While Intel indeed outperforms Arm based processors on a per/chip comparison, that's not the crux of the trade. What's more important is how it compares per a given dollar spent and the wattage consumed, depending upon a particular deployment. And Intel, being a near monopoly - technically speaking, a duopoly - has been known to muscle in on its customers. Arm, to the contrary, is an open market operation. This means that every manufacturer of laptops, desktops, tablets, servers and supercomputers will be motivated to escape from Intel's dominance. It may or may not happen instantly but that is the trend. Arm designed chips will require support from the developers, including Microsoft, and Windows 10 already work reasonably well on the Arm equipped Surface Pro. Samsung and Lenovo have Arm based laptops as well. But, thus far, it's been a tepid opening. Apple's switch from Intel is not tepid. It's the industry changing move.
 
You must log in or register to reply here.
Back
Top