Apple Says iPad Air Performance Details Are Accurate Despite 9-Core GPU Mistake

[bgillander]

Wow, this thread reminds me of a great SNL sketch:

 

[PaperMag]

😄 Do you think every mistake is a "typo" just because it's written down? I'm kind of shocked this requires explanation, but a typo is a "typographical error", an error in typography. It's essentially a transcription error introduced by the process of typesetting/typing it. With a typo, the author knows the correct information and accidentally transcribes that to incorrect text.

That is my point-- 9 to 8 would make a much more likely typographical error than 9 to 1 and certain 9 to 10. So no, I don't believe that's the most likely cause of the error. The more likely causes are miscommunications within the organization, lack of communication leading their technical writer to research for themselves how many cores are in an M2, or a change in specs after the original draft was complete leaving obsolete information to be published or perhaps to be resurrected when reverting to an earlier draft.




You have both said over and over, with certainty, that this is a typo. To wit:








(emphasis yours)


(emphasis yours)

Please retract your accusation that I'm lying, thanks.

We have said it was a typo. We never said we knew how a typo was made.

And I can’t believe I have to explain to this but here it goes:

A typo is the end result. How a typo was made—whether it was an accidental push of buttons, a lack of knowledge on how to correctly put down the written language, or the longer editing process by multiple stakeholders was lacking—its all a separate matter.

The point is someone was supposed to have put the correct number down, but they didn’t, and multiple stakeholders did not catch it before it went to “print.” Thus, the end result is a typo.

Again, I’ll repeat, the end result is not the same thing as the process by which the end result came to be. So using the term for one does not mean we know what happened. Just like using the term murder does not give us automatic insight into how the murder happened.

Ok, let’s stop talking. This is bottom of the barrel argumentation. We’ve hit rock bottom. It’s over.

 

[steve09090]

We have said it was a typo. We never said we knew how a typo was made.

The irony is, we as members on a forum can get a warning or a ban for correcting someone’s grammar, or a typo. But there is an entire story and 10 pages of banter based on Apple doing the typo.

Ok, let’s stop talking. This is bottom of the barrel argumentation. We’ve hit rock bottom. It’s over.

Yeah, 10 pages is enough.

 

[bgillander]

So, you will be paying for all those returned iPads? Internationally? Because, there is no returns for iPads (or any Apple product) after the initial 14-day period in my country. Just wanted to ask. No offense meant.

Even if Apple doesn't normally allow returns after 14 days in your country, I would expect one should be able to get a refund on this product due to it not meeting the advertised specifications. Since it more than meets the advertised performance, I'm not sure why one would want to unless they had a core fetish, but I expect Apple would not fight a return attempt due to their error, as the consumer has an obviously valid, specific complaint.

I could see this being a bigger issue if it were a Mac where there are actual options for increasing the core counts, but I personally think of the iPad more as a product where there is usually only a single model with just a few storage options... pretty much dealer's choice.

 

[PaperMag]

My simple question was whether Apple would accept it if a vendor delivers a product that is not meeting the specs. You are being very obtuse to avoiding saying that Apple would not accept it.
1. If Apple would be accepting stuff from vendors that is not meeting the specs, then Apple has a bigger problem.
2. If Apple does not accept stuff from vendors that does not meet the specs, then why should the consumer?

You're playing semantics and thus fooling the issue.

• Apple pays for chip specs because chip specs = 100% performance
• If Apple receives less chip specs, then they get < 100% performance
• Therefore either Apple still uses the chips (bins) or they have to not use it at all (trash)

In the case of customers, they are are receiving 100% iPad Air performance that was promised—"25 percent faster graphics"

We can stop arguing about this now. It's over.

 

[PaperMag]

Wow, this thread reminds me of a great SNL sketch:

Thats hilarious. My takeaway from the sketch is realizing everybody here is pretending to puke into their man purse.

 

[bgillander]

Thats hilarious. My takeaway from the sketch is realizing everybody here is pretending to puke into their man purse.

I just said it reminded me... I didn't say it was in a good way!

 

[PaperMag]

I could see this being a bigger issue if it were a Mac where there are actual options for increasing the core counts, but I personally think of the iPad more as a product where there is usually only a single model with just a few storage options... pretty much dealer's choice.

Thats exactly right.

• If Apple charges $100 for 25% faster GPU, and then under-delivers by 50% ← thats a problem
• But in the case of the iPad Air, Apple didn't promise 25% more GPU than an M1 model and then underdeliver ← Instead Geekbench scores prove customers in fact received 100% of promised performance

 

[bgillander]

Thats exactly right.

• If Apple charges $100 for 25% faster GPU, and then under-delivers by 50% ← thats a problem
• But in the case of the iPad Air, Apple didn't promise 25% more GPU than an M1 model and then underdeliver ← Instead Geekbench scores prove customers in fact received 100% of promised performance

Yes. I don't think it makes any sense to care more about the number of cores than the performance, which was accurately stated by Apple. That said, I do feel that if anyone really does care they should be able to return the iPad at this point, as it is only a few weeks into its lifespan (and should return it, rather than complain about some imagined slight). I cannot imagine Apple denying a return, regardless of their local policy, since Apple did make that mistake, but if they did deny a return, that would be the only way a lawsuit would seem legitimately possible to me.

Ironically, I expect that the Air does technically have 10 GPU cores, but only 9 are active, as that is the way binning usually works, simply disabling marginal cores.

 

[Beautyspin]

You're playing semantics and thus fooling the issue.

• Apple pays for chip specs because chip specs = 100% performance
• If Apple receives less chip specs, then they get < 100% performance
• Therefore either Apple still uses the chips (bins) or they have to not use it at all (trash)

In the case of customers, they are are receiving 100% iPad Air performance that was promised—"25 percent faster graphics"

We can stop arguing about this now. It's over.

If what you say is true (need citation that they only check the performance and not core count, whereas they clearly mention the core count on their (Apple's) sales page), then they (TSMC or the vendor) could put anything on the chip; they just have to make sure the performance is the same.

Can you see the flaw in your argument? It means, we cannot be sure of the core count of ANY of the chips delivered on ANY of the Apple's devices sold until now and probably far into the future.

I think you should cut your losses and agree that Apple dropped the ball and owe their customers an apology and refund/extend the return window.

 

[Beautyspin]

Even if Apple doesn't normally allow returns after 14 days in your country, I would expect one should be able to get a refund on this product due to it not meeting the advertised specifications. Since it more than meets the advertised performance, I'm not sure why one would want to unless they had a core fetish, but I expect Apple would not fight a return attempt due to their error, as the consumer has an obviously valid, specific complaint.

I could see this being a bigger issue if it were a Mac where there are actual options for increasing the core counts, but I personally think of the iPad more as a product where there is usually only a single model with just a few storage options... pretty much dealer's choice.

After 14 days, you are on your own in my country. They should extend the return window for these specific models and inform them that it is extended. After that, it depends on the individual.

 

[Beautyspin]

I’m interested to know how anyone can justify the number of cores (9 v 10) needed as a reason to return something. Other than that, it’s just being pedantic for the pure joy of 'getting one over the man'. Chest beating so to speak.

I have already given you the justification. Would Apple accept a shipment of 9 core GPUs when it had ordered a 10 core GPU without any hesitation?

 

[Beautyspin]

Yet, that’s a hypothetical because it did meet their specs. Because it was a spelling error in Advertising. Regardless of it being advertised as a 9 core, 10 core or even a 20 core, it performs exactly the same as it was benchmarked.

As far as a consumer choice, if a consumer 'required' 10 cores instead of 9, regardless of the performance then you may have a point, but surely it was the performance that was the critical point. What would you say if a person didn't like the advertised colour because it didn’t look exactly the same in their own environment?

So, you are confirming that the Apple QC department is clueless enough that it would accept a shipment of 9 Core GPUs when the order was for 10 core GPUs. See, I found the problem with Apple. They have a substandard QC department, if that were so. God only knows what else they have passed through their checks and when we will know about those snafus.

 

[Beautyspin]

That many GPU cores is usually for games that are close to AAA level or for image and video processing work.

That is not a number.

 

[PaperMag]

If what you say is true (need citation that they only check the performance and not core count, whereas they clearly mention the core count on their (Apple's) sales page), then they (TSMC or the vendor) could put anything on the chip; they just have to make sure the performance is the same.

Can you see the flaw in your argument?

You're not making any sense.

• Thats not how chips are made
• Thats not what I'm saying
• Your original analogy is not analogous, and therefor is invalid

I think you should cut your losses and agree that Apple dropped the ball and owe their customers an apology and refund/extend the return window.

I've said as much. I never argued otherwise.

This whole conversation, between me and you, was your error in reading comprehension and debate, you confusing the basics of how chips are made, and you confusing someone acknowledging a typo—for having full investigative awareness for how that typo was made—let alone you having clarity on the definition of the word "typo". Your choice strategy it seems is to lower the level of the conversation until your opponent gives up and I have to say you're quite effective.

 

[Beautyspin]

We've seen Apple accept subpar modems, inconsistent screens, Intel's processors for how many generations? Yeah, I think Apple would.

Without any complaint and for the same cost?

I think you're making this out to be more grave of an issue than it is. Apple is not perfect - no organization is. Mistakes happen, and the vast majority of the time it's something trivial, like this.

I think you are trivializing an issue that has far reaching consequences. What else have they missed?

Apple would notice right away if the delivered parts did not meet their specifications. This chip met the spec. Not sure why you keep pushing this idea that TSMC shortchanged them.

How can a 9 core GPU meet a 10 core Spec?

In the end it does not matter if a chip has 10 cores, 9 cores, 100 cores or 1 core if the performance is the same as advertised.

Why would it not matter if Apple would not accept a 9 core GPU for the same cost without demur? If they are accepting such products that do not meet the specs, what else have they been accepting and putting in their products?

If Apple had released a statement saying "all metrics in the presentation were taken from a 10-core GPU, but the shipped device only has a 9-core and does not perform the same" then yes, there would be a problem. But that's not what happened here.

They should have extended the return window. Period.

 

[bgillander]

After 14 days, you are on your own in my country. They should extend the return window for these specific models and inform them that it is extended. After that, it depends on the individual.

So what country is that exactly, as that is a truly crappy return window for Apple? I find it hard to believe that if the purchase was directly from Apple they would deny a return at this point in any country, considering their error, but I guess someone has to make the attempt before we will know for sure. If you actually have a new M2 Air and are this concerned, you would be much better off actually asking Apple about returning it, rather than kvetching on a forum, but if you have already tried that and been denied, then my apologies, and I believe you then do have an actual case against Apple. I personally feel it is an overreaction for a device that is performing exactly as it was represented, but if the core count is that important to you, as a customer you should have the ability to return it if you are unhappy that a specification was listed incorrectly at the time of purchase.

 

[bgillander]

I think you are trivializing an issue that has far reaching consequences. What else have they missed?

Sorry, but "what else have they missed?" is a really bad argument for "far reaching consequences". Apple got the performance specification correct, but they screwed up their search and replace in the technical specifications when they should have replaced the old M2 10 GPU cores with the updated 9 GPU core version. Apple are selling the consumer iPads, not CPUs, and if they misrepresent the device performance, that is a much larger concern than listing a technical spec incorrectly, even an embarrassing mistake like core counts.

 

[Beautyspin]

So what country is that exactly, as that is a truly crappy return window for Apple? I find it hard to believe that if the purchase was directly from Apple they would deny a return at this point in any country, considering their error, but I guess someone has to make the attempt before we will know for sure. If you actually have a new M2 Air and are this concerned, you would be much better off actually asking Apple about returning it, rather than kvetching on a forum, but if you have already tried that and been denied, then my apologies, and I believe you then do have an actual case against Apple. I personally feel it is an overreaction for a device that is performing exactly as it was represented, but if the core count is that important to you, as a customer you should have the ability to return it if you are unhappy that a specification was listed incorrectly at the time of purchase.

 

[Beautyspin]

Sorry, but "what else have they missed?" is a really bad argument for "far reaching consequences". Apple got the performance specification correct, but they screwed up their search and replace in the technical specifications when they should have replaced the old M2 10 GPU cores with the updated 9 GPU core version. Apple are selling the consumer iPads, not CPUs, and if they misrepresent the device performance, that is a much larger concern than listing a technical spec incorrectly, even an embarrassing mistake like core counts.

Are you telling us that Apple does not do QC for its components or it does not have the expertise to do so? I am not asking why Apple made the mistake. I am asking, "What would Apple have done if they were sent a shipment of chips with 9 core GPUs when they had ordered chips with 10 core GPUs." Would they have ketp silent, would they have asked for compensation, or would they have rejected the shipment?
This is an error that they caught internally. How did they catch it so late? Why was it not caught when the first iPad was manufactured and tested? Were they not tested? If they were tested earlier but the error was not caught, what changes did they make that caught this error. Why did they change the test procedure leading to the discovery. Would any more changes lead to more basic errors in other products.
How can you say it does not have far reaching consequences? If I (or someone) had told you a month ago that the iPad Air had 9 cores instead of 10, would you have believed? Who knows what else is missing in other products and whether they will be caught at all?

 

[PaperMag]

If I (or someone) had told you a month ago that the iPad Air had 9 cores instead of 10, would you have believed? Who knows what else is missing in other products and whether they will be caught at all?

Go look at how many GPU cores the $999 M2 MacBook Air has.

It’s going to blow your mind.

 

[MadDawg2020]

99.99% of the people buying the iPad Air could not care less about the core count. And even fewer would be able to notice a difference between the performance of 9 or 10 cores.
They buy it because it offers acceptable performance at a lower price.

The people who actually do care about such things - would be buying the iPad Pro.

 

[citysnaps]

We have said it was a typo. We never said we knew how a typo was made.

And I can’t believe I have to explain to this but here it goes:

A typo is the end result. How a typo was made—whether it was an accidental push of buttons, a lack of knowledge on how to correctly put down the written language, or the longer editing process by multiple stakeholders was lacking—its all a separate matter.

The point is someone was supposed to have put the correct number down, but they didn’t, and multiple stakeholders did not catch it before it went to “print.” Thus, the end result is a typo.

Again, I’ll repeat, the end result is not the same thing as the process by which the end result came to be. So using the term for one does not mean we know what happened. Just like using the term murder does not give us automatic insight into how the murder happened.

Ok, let’s stop talking. This is bottom of the barrel argumentation. We’ve hit rock bottom. It’s over.

I've finally come around... And now believe that Apple actually uses hot metal mechanical typesetting using a crew of men sitting behind Linotype typesetting machines to produce their documentation, as was done in the early 1900s.

Thus, using the word "typo" not only applies and is correct usage in today's common vernacular (a "mistake," as pointed out above), it's also accurate going back 100 years when the word was first used, and that Apple still uses. That should cover all bases.

 

[wbeasley]

OMG... this thread is like a Hydra....

Many many many people answer and explain it... and then one person pops up again and rehashes the whole thing and STILL doesnt get it ISNT about a lack of quality control and Apple being ripped off by the chip maker and not duping the purchaser of the iPad...

Aaaarrrrgggghhhhh.

I'm only going to keep watching to see if someone can make the penny drop for him... please.

 

[Analog Kid]

Once more… I never called you a liar.

Once more, I suggest you read the thread and realize I never said you did. Or just read the comment you replied to where I said I wasn't replying to you. Or the one before it where I said i wasn't replying to you.

This is who I was responding to:

Point to the comment where we purport to know that, otherwise stop saying untruths, thanks.

You might find yourself much less frustrated if you stopped assuming every comment here had your name on it and stopped assuming everyone you're talking to is the same person with the same point of view.