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

[Analog Kid]

Reduction from a number 10 to a number 9 in the GPU core count does not change the GPU performance they claimed in their advertisements or landing page.

They didn't claim a performance, they claimed a vague delta performance. I wouldn't be surprised if people were using the core count rather than the percentages Apple provided to estimate a performance number. Especially since the M2 family is well benchmarked already.

If Apple hadn't provided a core count to begin with, all would be fine. But they did, and it was wrong. You simply can not say that providing incorrect information is without consequence.

 

[ProbablyDylan]

I am not asking about perf. I am asking about cores. If TSMC delivers one core less, would Apple accept it?

TSMC would not make the chips with a core missing or disabled. They basically "print" the file that Apple provides. This was a decision Apple made.

More likely is a disconnect between departments. Engineering and Purchasing didn't accurately coordinate with Marketing, no one triple checked the slide before sending it. Mistakes like this sometimes happen.

 

[PaperMag]

They didn't claim a performance, they claimed a vague delta performance. I wouldn't be surprised if people were using the core count rather than the percentages Apple provided to estimate a performance number. Especially since the M2 family is well benchmarked already.

A vague delta performance is still a performance.

But I agree that some people could use the core count to assume it would get the same exact performance as the 10c chip in a MacBook Air, but even then Apple has underclocked chips in iPads before and so a discerning consumers should instead default to Apple's figures vs their own assumptions if they can help it.

It's a mistake none the less. Apple isn't excused from making this mistake twice.

 

[citysnaps]

A vague delta performance is still a performance.

But I agree that some people could use the core count to assume it would get the same exact performance as the 10c chip in a MacBook Air, but even then Apple has underclocked chips in iPads before and so a discerning consumers should instead default to Apple's figures vs their own assumptions if they can help it.

It's a mistake none the less. Apple isn't excused from making this mistake twice.

Apple isn't looking for or need to be excused.

Apple claimed its product offers customers a certain level of performance. And ships a product that does. From a customer's perspective it makes no difference if it has 1 or 100 cores.

I'm guessing 90% of Apple's iPad customer base have no idea what a "core" is.

 

[LogicalApex]

Apple isn't looking for or need to be excused.

Apple claimed its product offers customers a certain level of performance. And ships a product that does. From a customer's perspective it makes no difference if it has 1 or 100 cores.

I'm guessing 90% of Apples iPad customer base have no idea what a "core" is.

The decent thing to do is still to make it right. That would be at minimum letting affected buyers know and giving them an extended return window to compensate. Just in case it did matter to any of them. No matter how remote random forum members think that is.

That’s a customer first strategy.

 

[PaperMag]

Apple isn't looking for or need to be excused.

Apple claimed its product offers customers a certain level of performance. And ships a product that does. From a customer's perspective it makes no difference if it has 1 or 100 cores.

I'm guessing 90% of Apple's iPad customer base have no idea what a "core" is.

Agree with you 100% that in this instance no customer would or should care to return the product because Apple delivered 100% of the performance promised (my previous comment here testified to this). But lets be clear—they only get so much slack—they are not to make a habit of claiming one core count but delivering another.

 

[citysnaps]

Agree with you 100% that in this instance no customer would or should care to return the product because Apple delivered 100% of the performance promised (my previous comment here testified to this). But lets be clear—they only get so much slack—they are not to make a habit of claiming one core count but delivering another.

The decent thing to do is still to make it right. That would be at minimum letting affected buyers know and giving them an extended return window to compensate. Just in case it did matter to any of them. No matter how remote random forum members think that is.

Even when the overwhelming majority of Apple's customers have no idea what a "core" is?

And in consideration of the above... how about if during a keynote Apple claimed an upcoming product had a certain level of performance using a particular TSMC node, and then, for whatever reason, that changed to a slightly different node with the same level of performance and with zero adverse consequences.

Should Apple apologize for that and offer an extended return window?

 

[Analog Kid]

It's a mistake none the less. Apple isn't excused from making this mistake twice.

I don't think they're excused from making it once.

The question is what they should do to remedy it. The unit hasn't been on sale for very long, so the problem is relatively contained. I think they should be proactive in taking responsibility for the mistake.

Apple isn't looking for or need to be excused.

Apple claimed its product offers customers a certain level of performance. And ships a product that does. From a customer's perspective it makes no difference if it has 1 or 100 cores.

I'm guessing 90% of Apple's iPad customer base have no idea what a "core" is.

That's not all they claimed though. If they claimed it was midnight blue and instead it shipped as neon pink, people would have a reason to complain even if the performance was as promised.

They publicized a core number, beyond that they publicized it as the same configuration as the higher tier M2 MBA, which in itself gives it a premium appeal.

Apple can't argue that nobody cares about core numbers, because they don't bother advertising things that people don't care about. If they didn't think that the configuration was a selling point, they wouldn't have taken valuable space describing it.

 

[PaperMag]

Even when the overwhelming majority of Apple's customers have no idea what a "core" is?

And in consideration of the above... how about if during a keynote Apple claimed an upcoming product had a certain level of performance using a particular TSMC node, and then, for whatever reason, that changed to a slightly different node with the same level of performance and with zero adverse consequences.

Should Apple apologize for that and offer an extended return window?

Most iPad Air customers don't know what a core is, but Apple shouldn't get into the habit of mislabelling cores—obviously. What are we even arguing about?

iPad Air 9c vs 10c GPU is no big deal but it would very much be a big deal on the Mac side where stakes are higher and much more calculus is put into upgrade decisions, especially where GPU core upgrades is something Apple chargers for (unlike the iPad Air). Mislabeling specs is not a habit Apple wants to get into.

 

[citysnaps]

That's not all they claimed though. If they claimed it was midnight blue and instead it shipped as neon pink, people would have a reason to complain even if the performance was as promised.

Of course. If I ordered a blue iPad and received a pink iPad, then that would be a substantial meaning difference and not acceptable to me, and would return the iPad and get a refund. No doubt Apple would comply.

OTOH, if Apple claims a certain level of performance, and delivers an iPad meeting that level of performance (with zero adverse consequences; ie power consumption), I wouldn't care if it had just one core. Why? Because I'm not into feeling good about bragging how many GPU cores my iPad has.

 

[citysnaps]

Most iPad Air customers don't know what a core is, but Apple shouldn't get into the habit of mislabelling cores—obviously. What are we even arguing about?

I didn't realize Apple was into a habit of mislabeling cores. How many other times has that happened.

Imo... it's unreasonable to expect 100.0% perfection 100.0% of the time when humans are involved. Even for Apple.

No doubt in my mind that if Apple's typo resulted in even the slightest reduction in performance, they'd make it right in some manner - including a 100% refund if that's what a customer wanted.

 

[PaperMag]

I don't think they're excused from making it once.

So you don't forgive Apple for this mistake? They are forever in your ire and you'll never let it go?

The question is what they should do to remedy it. The unit hasn't been on sale for very long, so the problem is relatively contained. I think they should be proactive in taking responsibility for the mistake.

Apple should 100% remedy this should anyone want a return. It just won't happen at scale because in the real world people don't give as.

They publicized a core number, beyond that they publicized it as the same configuration as the higher tier M2 MBA, which in itself gives it a premium appeal.

Apple didn't advertise the core count in any of their promotional material. It was the tech specs page that had the typo. The psychological damage is much lesser than your tone suggests. And for what its worth the $599 M2 iPad Air with 9c GPU has one more core than the M2 MacBook Air at $999...

 

[Analog Kid]

Of course. If I ordered a blue iPad and received a pink iPad, then that would be a substantial meaning difference and not acceptable to me, and would return the iPad and get a refund. No doubt Apple would comply.

OTOH, if Apple claims a certain level of performance, and delivers an iPad meeting that level of performance (with zero adverse consequences; ie power consumption), I wouldn't care if it had just one core. Why?

Is there zero difference in performance between 9 cores and 10? You’re saying you looked at one spec and bought based on that. Others may have looked at a different spec and bought based on that. As I said, that 10 core M2 is a well understood part by now.

Because I'm not into feeling good about bragging how many GPU cores my iPad has.

Do you believe people who are should get shafted? That’s the audience that Apple is most likely to take brand damage from in this case.

 

[citysnaps]

Is there zero difference in performance between 9 cores and 10? You’re saying you looked at one spec and bought based on that. Others may have looked at a different spec and bought based on that. As I said, that 10 core M2 is a well understood part by now.

Your saying people make iPad purchase decisions based on core count rather than on performance specs? Weird. Because that wouldn't let a customer know what kind of performance to expect. Unless they were an engineer with intimate knowledge of the GPU core architecture, clock rates, integration with the OS, how many cores are used for a particular app, etc, etc.

 

[Analog Kid]

Imo... it's unreasonable to expect 100.0% perfection 100.0% of the time when humans are involved. Even for Apple.

So you don't forgive Apple for this mistake? They are forever in your ire and you'll never let it go?

If you make a mistake, you should own up to the consequences. Whether it's your first, our your umpteenth. That shows character.

You should hold others accountable for their mistakes. Remember that "first one's free" is a model meant to encourage the recipient to repeat that behavior as often as possible.

And let's not pretend that even those among us who have been very happy with Apple products over the years don't know this isn't Apple's first mistake.

 

[citysnaps]

Do you believe people who are should get shafted? That’s the audience that Apple is most likely to take brand damage from in this case.

Of course not. How would a customer get shafted if their new iPad performs as specified?

Seriously though... there may be some customers who will want to make a huge deal out of the typo, even though there are no adverse consequences with respect to expected performance.

No doubt in my mind Apple will offer a refund beyond the normal cutoff date to make them happy.

 

[Analog Kid]

Your saying people make iPad purchase decisions based on core count rather than on performance specs? Weird. Because that wouldn't let a customer know what kind of performance to expect. Unless they were an engineer with intimate knowledge of the GPU core architecture, clock rates, integration with the OS, how many cores are used for a particular app, etc, etc.

Such people don't exist? Have you not come across leman in these forums before?

The "performance spec" you're referring two are a general percentage rounded to the nearest 5%. It's meant to be fluffy, and anyone actually serious about performance is likely to dismiss it for exactly that reason.

I'll bet you I can find a workload that gets less of a boost than Apple's advertised numbers-- is that a reason to complain then? I'd argue no-- people should understand that performance varies with workload and no benchmark covers everything. But everyone also understands that 10 is more than 9, unambiguously.

If Apple didn't put 10 in writing, I wouldn't care either way. But they did. It was incorrect. They, and we, shouldn't sweep that under the rug.

 

[citysnaps]

If you make a mistake, you should own up to the consequences. Whether it's your first, our your umpteenth. That shows character.

You should hold others accountable for their mistakes. Remember that "first one's free" is a model meant to encourage the recipient to repeat that behavior as often as possible.

And let's not pretend that even those among us who have been very happy with Apple products over the years don't know this isn't Apple's first mistake.

I guess the solution, for you, is to find a tech manufacturer who lives up to that 100.0% level of perfection 100.0% of the time.

As I said previously, for those whose lives are driven by core count (rather than performance), there's no doubt Apple will make those people happy by refunding their money beyond the normal cutoff date. And then they can buy a Samsung iPad and find happiness.

 

[Analog Kid]

the typo

This isn't a typo. They didn't promote "9 corse". They promoted an incorrect specification: "10 cores".

 

[citysnaps]

Such people don't exist? Have you not come across leman in these forums before?

The "performance spec" you're referring two are a general percentage rounded to the nearest 5%. It's meant to be fluffy, and anyone actually serious about performance is likely to dismiss it for exactly that reason.

I'll bet you I can find a workload that gets less of a boost than Apple's advertised numbers-- is that a reason to complain then? I'd argue no-- people should understand that performance varies with workload and no benchmark covers everything. But everyone also understands that 10 is more than 9, unambiguously.

If Apple didn't put 10 in writing, I wouldn't care either way. But they did. It was incorrect. They, and we, shouldn't sweep that under the rug.

See my post above.

 

[PaperMag]

If you make a mistake, you should own up to the consequences. Whether it's your first, our your umpteenth. That shows character.

Nobody here is arguing otherwise. Apple made a typo, so they fixed the typo. And anyone harmed should return the iPad Air. But nobody is going to return the iPad Air. Because nobody was harmed.

 

[citysnaps]

This isn't a typo. They didn't promote "9 corse". They promoted an incorrect specification: "10 cores".

They specified 10 cores - a typo, as it should have been 9.

For those that are unhappy, they should simply return their iPad for a refund and find happiness from another tech manufacturer.

 

[Analog Kid]

I guess the solution, for you, is to find a tech manufacturer who lives up to that 100.0% level of perfection 100.0% of the time.

That's a completely disingenuous statement. I can't see how you would read my comments to suggest such.

there's no doubt Apple will make those people happy by refunding their money beyond the normal cutoff date.

Is there no doubt? I haven't seen it advertised anywhere. All I've seen is the damage control statement that went out. Until this becomes the policy you're sure it will become, there remains doubt. I'd like to see Apple compensate people who kept their units as well, even if it's a token gift card to say they recognize their mistake, but letting people know their return window has been extended would be a step in the right direction.

 

[citysnaps]

That's a completely disingenuous statement. I can't see how you would read my comments to suggest such.

I'm trying to help you/others understand that like people, companies occasionally make mistakes. Because perfection doesn't exist 100% of the time in the real world.

Simply take your iPad back to Apple for a refund and look elsewhere for a tech company with a better track record,

 

[Analog Kid]

Apple made a typo, so they fixed the typo

They specified 10 cores - a typo, as it should have been 9.

By "typo" you mean instead of typing "9", they accidentally typed twice as many keys, one of which on the opposite side of the keyboard?