October 18th Apple Event - Unleashed - Apple Silicon, MacBook Pro upgrades, HomePod mini, AirPods 3rd Generation

[leadeater]

1 hour ago, saltycaramel said:

Like, of course a Framework laptop is worse (or better: built with different priorities, priorities that just happen to make it worse for most people) if it does less work with more battery than an M1 Pro MBP. 

That's not a metric of why Framework is worse because it's repair and service focused, that's due to the hardware itself being less power efficient. If you want to make that conclusion then you have to compare to an equivalent laptop using the same hardware technology in a laptop that is not serviceable (or as serviceable).

 

What you said is not a statement about Framework at all, it has little to do with it at all in fact.

 

1 hour ago, saltycaramel said:

Even iFixit is kinda beginning to admit that there may be some correlation there. I loved the humility in their approach. “We still can’t say for sure”. They don’t have all the answers and they don’t automatically think they’re better than Apple’s engineers. The hearts of one thousand anti-integration r2r integralists when they said that maybe on such products soldered RAM and soldered flash modules shouldn’t be penalized in the score:

I don't think they included flash in their statement. CPUs/SoCs and DRAM failure is extremely rare so this certainly has a point, however NAND fails far more often so being soldered is still a negative mark towards reparability in a laptop.

 

You have to balance what the device is to be able to say what is a fair an reasonable compromise, because soldered NAND in a phone yea sure there probably isn't a better way without quite a significant compromise. However on a laptop, no matter how small can and will always be able to accommodate an M.2 device even if the NAND controller itself is in the SoC and not on the M.2.

 

Even with that said I'm actually rather confident a different mounting method could be designed for SoCs over the current BGA solder approach that is also better than LGA in a laptop device, it's just that there isn't an incentive or necessity to design such a thing. It won't become law and not saying it should but if it were such a solution would be designed quick smart.

[leadeater]

1 hour ago, Obioban said:

Is there a product category Apple plays in where their products don’t have the longest average service life? 

Business laptop sector, there's about no difference at all really. Consumer laptops are or can be massively inferior in that regard while Apple only sells a "single design" across their laptop product lines so it doesn't matter if you are a consumer or a business you get the same device with the same design.

 

HP/Dell/Lenovo really do need to do the same, it's just that they offer short warranties and not everyone bothers to use it when they fail and there are almost no incentives to try and get longer support or get it fixed by the manufacturer when it fails by the consumer. That's why consumer and business product lines get such a stark treatment difference, they more often make the failure of the device the manufacturers problem so they actually suffer from it.

[saltycaramel]

I had picked Framework as an example of “x86 non-integrated laptop” just because they’re the epitome of a laptop with a very desirable (for some) degree of serviceability, not because being from Framework has anything to do with the architecture having worse perf-per-watt than M1 Macs, but since I realize it can be misinterpreted I reworded that part this way:

 

Quote

Like, of course a very serviceable x86 laptop (without on package very-fast low-power LPDDR5 unified memory, without the SSD controller being a first class citizen on the main die and with potential for CPU-like year-over-year speed bumps, without the flash storage baseline speed being guaranteed by the laptop maker picking the flash modules besides creating and fine-tuning the controller, without everything that’s on package being pretty close to each other which as we know generally is a good thing, etc.) could be proven (for many usages, scenarios and most people) to be worse (or to say it better: NOT WORSE IN GENERAL, just built with different priorities, priorities that just happen to make it worse for most people) IF it does less work while using more battery juice than an M1 Pro MBP or an M1 MBA. 

 

[HenrySalayne]

2 hours ago, saltycaramel said:

3) Android/iPad tablets and smart TVs are pretty large, why can’t I upgrade the main boot storage and the RAM?

Ok, fair question, but I got you covered:

M.2 drives would be quite hard to fit into a phone. But if you reach something in the neighbourhood of 10 inches of display size, there is always enough space to fit an M.2.  (and Microsoft proved that with their newer Surface Pro). The level of integration on phones is a necessity, on larger devices it is a choice.

2 hours ago, saltycaramel said:

4) phones have waaaay more Z-space for webcams than laptop lids because they’re thicker than any laptop lid

It's not "waaaaay more" it's maybe one or two Millimetres. And if you disassemble phones and notebooks you might find the front facing camera modules to be quite similar in size. But there is even another solution: you can build an indentation into the chassis of your notebook. I hear you say "but this is madness!" to which I can only reply that the clever folks at Apple have that covered. There is already a indentation in the right spot but it's currently only used so our fat fingers might squeeze more easily under the lid.
If we accept camera bumps on the back if phones we can certainly accept camera bumps at the top of a display. And I would see something like this much rather than a notch (or a really bad camera).

Or: just take an array of bad micro cameras across the top and use some clever image processing to compile an image with less noise and more details. Which is also already done by almost any company building phones. You could also get depth information for face recognition. This would be innovative - a gigantic notch or a bad webcam (or none whatsoever) is just the lazy way to do it.

 

[saltycaramel]

I ask if upgrading RAM is a common thing in the general population even in upgradeable laptops.

Answer: “used to be thing back in the day”

 

I ask if upgrading the SSD is a common thing in the general population even in upgradeable laptops.

Answer: “not really, but..”

 

I ask if it is a common thing for consumer electronics (pricey smartphones, mobile-OS tablets, pricey smart TVs with extremely long lifetime, etc.) to be able to upgrade the boot drive and RAM

Answer: “phones are small (yeah, like there wouldn’t be a way to concoct a fast&reliable microSD-sized thingy to be at least professionally serviceable), tablets not really but that one time Microsoft proved it can be done but it’s not really a thing for other tablets”

 

 

People accidentally proving my point. 

 

 

As for NAND being more susceptible to failures than RAM, it’d be interesting to know how many iPhones, iPads and post-2016 Macs ended up dead because that particular kind of failure..to gauge how big of an actual problem it is…

[Obioban]

3 hours ago, leadeater said:

Business laptop sector, there's about no difference at all really. Consumer laptops are or can be massively inferior in that regard while Apple only sells a "single design" across their laptop product lines so it doesn't matter if you are a consumer or a business you get the same device with the same design.

 

HP/Dell/Lenovo really do need to do the same, it's just that they offer short warranties and not everyone bothers to use it when they fail and there are almost no incentives to try and get longer support or get it fixed by the manufacturer when it fails by the consumer. That's why consumer and business product lines get such a stark treatment difference, they more often make the failure of the device the manufacturers problem so they actually suffer from it.

Disagree— most of those MacBooks get a second life as personal computers. E.g. my wife’s computer is a 2012 15” MacBook Pro that spent the first 5 years of its life as a corporate issued laptop at the hedge fund my friend works at.

[leadeater]

2 hours ago, saltycaramel said:

I had picked Framework as an example of “x86 non-integrated laptop” just because they’re the epitome of a laptop with a very desirable (for some) degree of serviceability, not because being from Framework has anything to do with the architecture having worse perf-per-watt than M1 Macs, but since I realize it can be misinterpreted I reworded that part this way:

Well there is a couple of problems here.

 

• Unified Memory Architecture doe not necessitate on package DRAM (or LPDDR)
• SSD controller within the SoC or on the M.2 itself has no effect on it's performance, latency or bandwidth. Or more correctly the difference is so miniscule there would be no software tools with the require accuracy to measure it.
•  Having everything on package is not automatically better and neither has much benefit as you'd believe.

 

On the first point Apple neither pioneered this or even perfected it. Both Nvidia and AMD have being doing Unified Memory SoC's and designs for a very long time, AMD in consoles and Nvidia in ARM Socs for mobile devices and automotive industry, This is also restricting the UMA discussion solely to shared physical memory implementations as UMA is not restricted to this specifically, UMA is part of both Nvidia CUDA and AMD's w/e the hell they call theirs now. UMA with CUDA/OpenCL however is most useful when also utilizing NVLink/Infinity Fabric (GPU to GPU).

 

The reasoning for why Apple went with on package LPDDR5 is because there is no practical way to have enough SoC pins and motherboard traces to do a 256bit wide or 512bit wide memory bus. The key factor here for why is physical space not latency or power of having it so close to the CPU/SoC or even on the package.

 

This the same reason HBM is also deployed thus far on package because that is 1024bit or wider, a simple 2 HBM package configuration requires a 2048bit memory interface. That is why every GPU that has used HBM has used it on package. The widest GDDR implementation used was way back on AMD 290X/390X series cards and that was very challenging to get that on there.

 

 

Something like this just is not practical for a CPU in a laptop that also needs to provide PCIe lanes and chipset connectivity. Even on GPUs nobody attempted 512bit wide GDDR ever again, board design is too complicated and expensive.

 

So to sum all that up on package LPDDR is not a power, performance, latency or bandwidth benefit for either Apple or anyone else save for the fact it is the only practical way to have such a wide memory interface which is a physical aspect issue.

 

For the SSD point, really not much to cover here. The only real benefit to have the SSD controller in the SoC itself is the power efficiency and reliability of these and the ability to use the DRAM of the SoC rather than dedicated DRAM on a M.2 or SATA device.

 

On the whole general on package benefits this is in general vastly misunderstood and doesn't give any or as much benefit as believed. The degree of power savings for having it on package versus close by is minor. The latency difference is immeasurable and other factors are order of magnitude greater so the difference would be entirely lost due to those.

 

Having things on package has significant benefits in design simplicity and making things that were otherwise impractical practical.

 

Looking at any of the Apple laptop SoCs and comparing to any x86 laptop, especially a Framework one, and trying to draw any conclusions about the power efficiency benefits of having things in SoC or on package is extremely flawed and at worst misleading.

 

Fortunately we do have such examples in the x86 laptop sector, Intel Tiger Lake laptops have both on package LPDDR4X and off package LPDDR4X and I'll save you the hassle of having to look through enough reviews to get a consistent idea of the difference, none.

[leadeater]

3 minutes ago, Obioban said:

Disagree— most of those MacBooks get a second life as personal computers. E.g. my wife’s computer is a 2012 15” MacBook Pro that spent the first 5 years of its life as a corporate issued laptop at the hedge fund my friend works at.

So do retired business laptops, they get handed on just the same and there are even dedicated companies refurbishing them and selling them.

 

https://www.pbtech.co.nz/category/computers/exleased/laptops

 

So why is it you think that's unique to Apple? It's not.

[leadeater]

1 hour ago, saltycaramel said:

I ask if upgrading the SSD is a common thing in the general population even in upgradeable laptops.

Answer: “not really, but..”

One thing I would say is HDD to SSD upgrades were a lot more common, not so much anymore since only the worst trash comes with HDDs now heh

 

1 hour ago, saltycaramel said:

As for NAND being more susceptible to failures than RAM, it’d be interesting to know how many iPhones, iPads and post-2016 Macs ended up dead because that particular kind of failure..to gauge how big of an actual problem it is…

I would say quite low since the workload from these devices is quite low on the NAND and also Apple doesn't equip them with budget tier low end NAND either, unlike many Android devices, particularly cheap tablets. Those I have seen NAND failures.

[igormp]

8 minutes ago, leadeater said:

UMA with CUDA/OpenCL however is most useful when also utilizing NVLink/Infinity Fabric (GPU to GPU).

But that's only virtual (nvidia even calls it like that), since you're still constrained by the bus bandwidth and need to work around that. On a real UMA setup you don't need to care at all about that.

 

Other than that, they do have actual UMA devices, so your point still stands anyway.

[leadeater]

1 minute ago, igormp said:

But that's only virtual (nvidia even calls it like that), since you're still constrained by the bus bandwidth and need to work around that. On a real UMA setup you don't need to care at all about that.

It's still real UMA, UMA doesn't require sharing the same physical memory and isn't the sole benefit of UMA. But yes that's why I said you really need NVLink GPU to GPU otherwise it being in CUDA is basically worthless.

[saltycaramel]

So to recap, there are some things that could actually be done differently with little to none impact and 1-2 crucial things (512bit wide bus, efficient/sustainable manufacturing) that make this the only practical way to do it.

 

(I’d note that there may also be some other technical reasons pertaining the two multi-chiplet products we haven’t seen yet, the M1 Max “Duo” and the M1 Max “Quad”)

 

We’re left with one thing to do: follow the iFixit approach and wait&see if someone else in the coming years will catch up (in many ways, be it performance-per-watt, graphical power in compact designs, memory bandwidth, etc.) to this kind of design but by using a more modular/serviceable design. For sure someone will pick up the gauntlet, if it’s possible. 

 

[HenrySalayne]

3 hours ago, saltycaramel said:

I ask if it is a common thing for consumer electronics (pricey smartphones, mobile-OS tablets, pricey smart TVs with extremely long lifetime, etc.) to be able to upgrade the boot drive and RAM

Answer: “phones are small (yeah, like there wouldn’t be a way to concoct a fast&reliable microSD-sized thingy to be at least professionally serviceable), tablets not really but that one time Microsoft proved it can be done but it’s not really a thing for other tablets”

People accidentally proving my point. 

Do they? Just a friendly reminder:

23 hours ago, saltycaramel said:

That’s definitely different from r2r “purists” who refuse to see the advantages of some integrated solutions and fail to explain why some things are acceptable on smartphone/tablets but not laptops. 

M.2 drives are a reality and widely available. On the other hand, micro SD cards are great for bulk storage but suffer in IOPs and access times. There is no comparable small format equivalent for M.2 drives.

You cannot blame phone manufacturers for being unable to built M.2 slots into their devices. But it's quite easy to implement exchangeable storage in larger devices. Like Microsoft did on the Surface without compromising size, speed, battery life, design or anything else. That is the point here.

 

And what was your point again? And how did I accidentally proof your point?

[Obioban]

1 hour ago, leadeater said:

So do retired business laptops, they get handed on just the same and there are even dedicated companies refurbishing them and selling them.

 

https://www.pbtech.co.nz/category/computers/exleased/laptops

 

So why is it you think that's unique to Apple? It's not.

No, my point was that your statement of businesses only use them for the duration of the warranty isn’t meaningful if their life doesn’t end when the businesses retire them.

 

And, failing that, they fall into the overall category of laptops, of which Apple’s do have the longest average service life.

 

Which is what I’m trying to get here— people are complaining that Apple’s construction methods are making e-waste, but their phones, tablets, computers all have the actual longest average life in the industry. So, while you may disagree with the details, they’re actually getting the results you’re wanting better than anyone else.

[leadeater]

14 minutes ago, saltycaramel said:

So to recap, there are some things that could actually be done differently with little to none impact and 1-2 crucial things (512bit wide bus, efficient/sustainable manufacturing) that make this the only practical way to do it.

I would actually add a 3rd, the SSD controller in the SoC does have a large reliability benefit so you know for sure it's of the utmost highest production quality standard and will have very well designed power delivery and protection. Also performance wise that controller is going to be very good as well. I'd just like to said that paired with replaceable NAND storage on something similar to or on an M.2. I really do think 3-5 years down the line it would be very nice to be able to upgrade that storage to the latest generation while not having to pay such a high premium on equivalent capacity. In the past I wouldn't have said this would be needed all that much but with how good video editing is I really do think more people are going to be using MacBook Pros for this than ever before and on much higher quality source content.

 

Might even be something Apple adds to the next generation if enough ask for it, or add an additional expansion capability. I am thinking however it's a big enough change Apple couldn't/wouldn't do it next year under a minor model year refresh.

[leadeater]

12 minutes ago, Obioban said:

No, my point was that your statement of businesses only use them for the duration of the warranty isn’t meaningful if their life doesn’t end when the businesses retire them.

That's not what I said nor what you stated that I was replying to. You said is there any category Apple is in that they do not have the longest service life, my answer to that is business laptops where I rate these as equal.

 

12 minutes ago, Obioban said:

And, failing that, they fall into the overall category of laptops, of which Apple’s do have the longest average service life.

No they are equivalent only. Handing on replaced business laptops is very common, this isn't an "Apple" thing.

 

Business laptops are not consumer laptops, these are entirely different product lines. It doesn't matter how many garbage consumer laptops for example HP makes it has literally no bearing at all on the service life of their business products.

 

This is as I said a key benefit to what Apple is doing, there is no consumer product line and business product line, it's all the same which is a good thing.

 

Edit:

Maybe you just didn't understand the point about warranty and support, because business laptops become the problem of those who make it if they do a bad job so they actually make them better so they do not become their problem which means they last longer. $10-$15 extra cost in design, manufacturing and parts could literally save them hundreds of thousands to millions of dollars in on site warranty repair and parts replacements. There actually is a direct financial benefit to designing a better and more reliable laptop in the business sector.

[saltycaramel]

10 minutes ago, HenrySalayne said:

And what was your point again? And how did I accidentally proof your point?

 

My point (by asking how frequent/important/deal_breakers some upgrade procedures are out there in the real world) is that maybe some vocal tinkerers are the 5% demanding design trade-offs that benefit that 5% and don’t necessarily benefit the 95% (that maybe is better served by preposterously good speakers, beefier cooling and the motherboard PCB getting smaller and smaller and smaller).

 

On the other hand I didn’t shy away from saying that the way Apple manages to put psychological pressure on you (I just went thru this) when you’re on that CTO configuration page and you feel compelled to future proof your RAM and storage choice is almost an art at this point. So there’s that too. 

[leadeater]

2 minutes ago, saltycaramel said:

the motherboard PCB getting smaller and smaller and smaller

What benefit is there really in say a 14" laptop if the main board takes up 30% instead of 40%. Battery capacity is capped at 99Wh and then you are on diminishing returns for things like speakers quite quickly. Having a smaller mainboard does have a benefit but it's not a linear benefit as the size decreases, at some point you actually gain nothing other than empty volume. I don't personally know what that point is because I think that largely depends on how much power and thus cooling the CPU and GPU, or SoC, needs.

 

If Apple were still using Intel i9's with peak power in the 150W+ range than I'd definitely see a smaller mainboard as a huge benefit, being able to jam in a MASSIVE cooling solution lol.

 

In an indirect way Apple solved the need for more internal volume in laptops with the M1 family of SoCs by making them so efficient existing cooling solutions are well above what is necessary. Lets just say for example sake Apple has been able to make the mainboard 10% smaller, well they can make the cooling solution literally half the size 

[Kisai]

15 minutes ago, leadeater said:

That's not what I said nor what you stated that I was replying to. You said is there any category Apple is in that they do not have the longest service life, my answer to that is business laptops where I rate these are equal.

 

No they are equivalent only. Handing on replaced business laptops is very common, this isn't an "Apple" thing.

Used car analogy.

 

There are two kinds of pre-owned cars. Off-lease, and customer-trade-in.

Off-lease vehicles have high mileage but are otherwise well maintained (in theory.)

Customer trade-ins are usually just cleaned up, any safety problems fixed and put on the lot. 

 

Most of the time customer trade-ins are given new life as those $1000-$10,000 off-warranty cars that become "teen's-first-car", and they use it until it becomes a safety-hazard to themselves. Off-lease vehicles however, might not have even run out of their original warranty mileage if the vehicles were mainly used in-city.

 

But with electronics, most electronics end up in the landfill. Smartphones, tablets and televisions almost never see new life, because they have no means for the end-user to get any life out of them. 3 years old? no more updates. A recent "device breaking" update was the root certificates update for lets encrypt. So if you are using a device that hasn't received any updates since 2018 (such as a smart tv's and ipads) you're going to start seeing a lot of "broken" sites that won't work, because the site or it's frameworks/libraries linked to CDN's won't work. Thanks for breaking the internet Google.

 

With cars, the only thing that might keep you from using a 3+ year old car, are emissions testing in states that have it. Which scummy people will just go have their vehicle "fixed" to pass the test, and the restore it to the failing emissions to get the performance back. Remember "cheating emissions" was thing that several car manufacturers themselves got in trouble over. As we move to EV-only policies, there will come a point where you won't be able to use an old ICE vehicle, and millions of people will be stuck with vehicles they can't trade-in, off to the junk yard they go, despite still being functional.

 

That's the same point we're reaching with laptops. Everything gets soldered to the PCB, thus making reuse/resellability difficult as there's no way to guarantee the device won't fail immediately. It would be nice to get 10 years out of a laptop, but only Apple even commits to having hardware last 7 years, most other laptop hardware, the battery is done in 2 years, and the cooling fans are done in 1. PCB? Perfectly fine. But you can't get the parts a year after the warranty expires. Dell/HP/Lenovo doesn't make that many spares, they just refurbish stuff until they're out of parts and then sucks-to-be-you.

 

 

[Obioban]

52 minutes ago, leadeater said:

That's not what I said nor what you stated that I was replying to. You said is there any category Apple is in that they do not have the longest service life, my answer to that is business laptops where I rate these as equal.

 

No they are equivalent only. Handing on replaced business laptops is very common, this isn't an "Apple" thing.

 

Business laptops are not consumer laptops, these are entirely different product lines. It doesn't matter how many garbage consumer laptops for example HP makes it has literally no bearing at all on the service life of their business products.

 

This is as I said a key benefit to what Apple is doing, there is no consumer product line and business product line, it's all the same which is a good thing.

 

Edit:

Maybe you just didn't understand the point about warranty and support, because business laptops become the problem of those who make it if they do a bad job so they actually make them better so they do not become their problem which means they last longer. $10-$15 extra cost in design, manufacturing and parts could literally save them hundreds of thousands to millions of dollars in on site warranty repair and parts replacements. There actually is a direct financial benefit to designing a better and more reliable laptop in the business sector.

I just don't see how any of it matters-- if the laptop's life isn't over when it's done being a business computer, it doesn't matter than the business stops using them at the end of the warranty period. The lifecycle is done when it becomes e-waste. 

[leadeater]

13 minutes ago, Obioban said:

I just don't see how any of it matters-- if the laptop's life isn't over when it's done being a business computer, it doesn't matter than the business stops using them at the end of the warranty period. The lifecycle is done when it becomes e-waste. 

Yes and if both get used after the business or whoever first purchased it and it goes on to another person to use then which is being used longer? Both right?

 

So then which has the longer service life, neither....

[igormp]

2 hours ago, leadeater said:

It's still real UMA, UMA doesn't require sharing the same physical memory and isn't the sole benefit of UMA. But yes that's why I said you really need NVLink GPU to GPU otherwise it being in CUDA is basically worthless.

No, it's not. UMA means that both devices can access the same memory address with the same latency without any abstraction.

 

When you add abstractions such as CUDA or other libraries in between in order to make such accesses across devices easier, that's called virtual or hybrid UMA, but is only an abstraction to make it easier for people to work on without any of the performance benefits.

 

Even when you get to large CPUs it's often hard to do UMA, and you can only do NUMA mode.

[leadeater]

25 minutes ago, igormp said:

No, it's not. UMA means that both devices can access the same memory address with the same latency without any abstraction.

 

When you add abstractions such as CUDA or other libraries in between in order to make such accesses across devices easier, that's called virtual or hybrid UMA, but is only an abstraction to make it easier for people to work on without any of the performance benefits.

 

Even when you get to large CPUs it's often hard to do UMA, and you can only do NUMA mode.

I think you are mixing terms, this is true for Uniform Memory Access (UMA) not Unified Memory Archecture (UMA). To be fair UMA is not a common short name or acronym for Unified Memory Archecture but I couldn't be bothered typing it in full more than once.

 

UMA vs NUMA actually is not related to Unified Memory at all and Unified Memory can and is used in NUMA systems. Every multi socket system is a NUMA system and that may have 4, 8 or 16 Nvidia GPUs with GPU to GPU NVLink with the CUDA application utilizing Unified Memory Model.

 

Quote

With CUDA 6, NVIDIA introduced one of the most dramatic programming model improvements in the history of the CUDA platform, Unified Memory. In a typical PC or cluster node today, the memories of the CPU and GPU are physically distinct and separated by the PCI-Express bus. Before CUDA 6, that is exactly how the programmer has to view things. Data that is shared between the CPU and GPU must be allocated in both memories, and explicitly copied between them by the program. This adds a lot of complexity to CUDA programs.

 

Unified Memory creates a pool of managed memory that is shared between the CPU and GPU, bridging the CPU-GPU divide. Managed memory is accessible to both the CPU and GPU using a single pointer. The key is that the system automatically migrates data allocated in Unified Memory between host and device so that it looks like CPU memory to code running on the CPU, and like GPU memory to code running on the GPU.

Unified Memory is a management model, not a characteristic of memory access patterns (latency and bandwidth uniformity) or the physical sharing of memory itself.

 

Edit:

On say an Intel CPU with an iGPU that is sharing system memory this would be Uniform Memory Access for both the CPU and GPU however it's not Unified Memory. That is also true for AMD APUs in PCs too a believe. Apple M1 SoCs is taking advantage of both Uniform Memory Access and Unified Memory so when a pointer is used by either the CPU or GPU no memory copy operation is required.

 

iGPUs and APUs in PCs when they share the system memory the memory that is allocated to the GPU is virtually partitioned and treated as if it were physically different memory and the CPU can no longer use that portion of system memory. Memory copy operations are therefore required, which means data is moved from the memory to the CPU and then back to the memory.

[Obioban]

29 minutes ago, leadeater said:

Yes and if both get used after the business or whoever first purchased it and it goes on to another person to use then which is being used longer? Both right?

 

So then which has the longer service life, neither....

I assume they both get factored into overall computer longevity, which is an area where apple leads.

[leadeater]

9 minutes ago, Obioban said:

I assume they both get factored into overall computer longevity, which is an area where apple leads.

Yes but you are stating an opinion that you think they last longer, have a longer service life. If you however break it down in to market segments then that tells a very different story and highlights quite a big problem in the industry, a problem that like I said Apple does not share.

 

If you are talking about consumer laptops then I agree, if you are talking about business laptops then I do not agree.

 

You asked where Apple does not lead, business laptop market where they are all very much equal. Why is this difficult to understand?

 

An HP EliteBook is not equivalent to or even in the same market category as an HP Pavilion.

 

Consumer laptops and business laptops are district different markets, they shouldn't really be but they are. Apple sells the exact same devices in to both markets.

 

They are different even down to the CPUs that are in them, business laptops come with Intel CPUs that support things like Intel vPro.