AMD's new naming scheme for its mobile CPUs seems purposefully confusing

[Redaktion]

AMD has announced that next year's mobile processors will use a new naming scheme. The new system is difficult to understand and may confuse customers - maybe on purpose.

https://www.notebookcheck.net/AMD-s-new-naming-scheme-for-its-mobile-CPUs-seems-purposefully-confusing.648985.0.html

[Bareback]

it's like 4 digits
You'll manage

[watcher]

How about looking at this from a different perspective:
If AMD creates the entire 7xxx product line specifically with "more means better" in mind, when looking at the numbers, it could really line up with average performance increase and price increase. We don't known this yet, do we?
I bet 99% of people looking at those model numbers have no idea how it works between each generation, and probably don't want to either. It is possoble that it could be more confusing precisely for the remaining 1% who do know all this. So which one is more important in the end? 99% i 1%?
Time will tell...

[_MT_]

If AMD creates the entire 7xxx product line specifically with "more means better" in mind, when looking at the numbers, it could really line up with average performance increase and price increase. We don't known this yet, do we?

Yes, most people have no clue and would probably assume that higher number is better. Fundamentally, the problem is that when you have different architectures, different manufacturing nodes, different bins, different numbers of cores, which processor is better depends on your workload. It's impossible to name a diverse group so that higher number is always better. And that's before we consider things like accelerators for video decoding. You can have a Ryzen 9 processor than has the same number of cores as a Ryzen 7 processor where the only difference really is binning and perhaps power limits. But what if the Ryzen 7 unit is build on architecture that is two generations newer? Or course, it seems unlikely that they would sell old architecture in a high-end line that is Ryzen 9. But placing line-up, a marketing construct, as more important than architecture, that doesn't sit well with me.

What I want to know as a customer is number of cores, architecture, manufacturing node and perhaps binning (although I'm usually reluctant to pay for high bins as they tend to be poor value). Traditionally, series would indicate architecture and manufacturing node. If you reduce series to mean just a model year, it becomes practically useless. The second number designates number of cores but not directly - it sometimes designates just a higher bin. And while binning makes a difference, it's a different kind of difference than number of cores. You could debate whether number of cores is more important than generation of architecture, it would depend on your workload and how big the generational gap is, but binning in my book definitely isn't more important than generation.

If they want to sell older Ryzens, they should sell them under older series or under different name like Athlon. They could have Athlon lag behind Ryzen by two generations for example. I know that it's not desirable from marketing standpoint. That current generation Ryzen is what people want. But they're just diluting what "current generation Ryzen" means. The last thing you want is for people to have high expectations of a product then being disappointed. If reviewers get the newest stuff and people unwittingly buy the older stuff, you're playing with fire. AMD has a history of screwing this up.

[NikoB]

All she does is go round and round she goes. Where she stops, nobody knows (We want twist technically illiterate consumers)...
The actions of large TNCs inevitably come down to natural fraud...

Experts will figure it out anyway, and illiterate townsfolk, as always, will rely on their opinion and blind faith. Amen!

[watcher]

Yes...

I get you, you bring valid points. You and I both belong to the 1% that cares about this kind of details.

As I understand, AMD as a business in that kind of market fundamentally has to care about the whole market or at least it's majority. Rules of the market. I would be surprised if they didn't do  market research to try to figure out how to optimize this, to find a reasonable compromise that works best for everyone. Maybe this really is a genuine research result.

Try to think like an average Joe. Imagine that you have no idea how to choose between 10 different products. Would say you feel uneasy, uncomfortable with making a decision? The problem is the paradox of choice. Your more likely to not choose anything at all or at least delay the decision. They still need to try to sell.

Aside from that theres the problem that manufacturers often configure the same cpu model with different cooling, and you get vastly different results in laptops that barely look different from each other. AMD can't force anything, without actually trying to label cpus differently per different purpose.

Look what Apple does with their lineup.
Then look how crazy Arm based mobile cpu market looks like. Its even worse.

[Bau]

Bifore to make a ot of considerations it is worth to wait and see...
I mean, how you can know now that a higher number does not mean a better cpu? For example you could have:

7640 -> Speculation: 6-Core Zen4 @ x-Ghz (i.e., actual 7000gen Ryzen 5 Zen4)
7730 -> Speculation: 8-Core Zen3 @ x-Ghz (i.e., equivalente to 6000gen Ryzen 9 Zen3)
7740 -> Speculation: 8-Core Zen4 @ x-Ghz (i.e., actual 7000gen Ryzen 7 Zen4)

Moreover you could have such difference only for U/H/HX and you may not have some model, for example no Zen 3 iteration for 78xx.

In conclusion we will see, it may happen that the 95% of people who do not know the maining of the number will be happy with the choise "the bigger, the better"; the 4% that know the maining of the number can make a better choice and the other 1% do not need any explanation about the name because they will search exactly that cpu. Or can be a disaster.

[Russell]

Well.
AMD explained it's nomenclature for the sake of people who might actually care.
And this explanation is going to be available to view at any point of time later.
Those who don't care are still more likely to be cheated and persuaded to by a kaybylake 'i7' than a 'Ryzen 5 7xxx' or Alderlake or newer i3.

Such information wouldn't be of any benefit to those kind of customers no matter how much clarity is provided.

Besides, these assumptions about the first two indicating architecture are not really some axioms.

To me, the only issue is that it doesn't say anything about the iGPU.

The issue of having an older architecture in the same Ryzen 7 or Ryzen 5 is quite likely to be exploited by sellers in amazon or best buy.
They did it with Lucienne, no surprises if they do it again.
This is the real issue here.
If they actually provide the model number rather than just 'configurable with up to Ryzen 7 or 9' then there is no issue.

[Erik]

They did it with Lucienne, no surprises if they do it again.

It's no surprise because AMD has enabled them to do it again, I'm with Mr Herzbig on this point, AMD reversed its stance on this confusing approach with Barcelo, which remained part of the Ryzen 5000 series albeit it was launched together with Rembrandt. It's likely that it didn't work very well with AMD and the OEM if they're now returning to the tradition of confusing costumers, with the spin that now under the Ryzen 7000 you'll find not two, not three, but four Zen generations (Zen 2, Zen 3, Zen 3+, Zen 4). Even Intel didn't reach this level of wickedness when it comes to naming, but this just shows that it didn't take AMD long to surpass the master.

[Russell]

They did it with Lucienne, no surprises if they do it again.

It's no surprise because AMD has enabled them to do it again, I'm with Mr Herzbig on this point, AMD reversed its stance on this confusing approach with Barcelo, which remained part of the Ryzen 5000 series albeit it was launched together with Rembrandt. It's likely that it didn't work very well with AMD and the OEM if they're now returning to the tradition of confusing costumers, with the spin that now under the Ryzen 7000 you'll find not two, not three, but four Zen generations (Zen 2, Zen 3, Zen 3+, Zen 4). Even Intel didn't reach this level of wickedness when it comes to naming, but this just shows that it didn't take AMD long to surpass the master.

Barcelo is a different case. If I am not mistaken, all the processors under 6xxx series at least got RDNA instead of Vega.
It would caused a lot of backlash had they just named it under 6xxxx series.

Intel's naming doesn't matter, i7/i5/i3 are the best is something imprinted in the brains of layman due to years of marketing.
They can still sell kabylake i7 laptops with (it has an i7, but 50% off) and get away with it, because people who actually care would explore deeper and those who don't would be happy with something that just works and helps them browse through webpages.

Though I do agree with you on the issue of possible exploitation by sellers, I think this 'standardized' naming scheme is actually pretty good. At least better than just saying 'this time, there are both Zen 2 and 3 under the same Ryzen 7 chips'.

Though it is indeed like saying 'We've explained our stand, you know what to look for. In case you make a mistake, it's not our fault'.
But then again, shouldn't the buyer hold some accountability?
We have so many reviewers these days, even googling would help a little if done properly.
Again, I'd have liked it had the nomenclature included info on the gpu cores as well..

[LostInSpace]

Wow, this makes Intel and Dell XPS numbering look 'sane' by comparison. (Just kidding!, Everyone's numbering is insane!)

I5's started with no multithreading, then they got that too!

XPS 15 9530 (2013) - OK
XPS 15 9550 (2015) - OK
XPS 15 9560 (2017) - OK
XPS 15 9570 (2018) - OK, so far so good
XPS 15 7590 (2019) - What???
XPS 15 9500 (2020) - What??? Back to the future??? See 2013....
XPS 15 9510 (2021) - OK, back in sequence, I guess....
XPS 15 9520 (2022) - OK, but what are you going to do next year, when you land on 2013's model number?

Ha, ha, ha, ha.....

[Russel]

Wow, this makes Intel and Dell XPS numbering look 'sane' by comparison. (Just kidding!, Everyone's numbering is insane!)

I5's started with no multithreading, then they got that too!

XPS 15 9530 (2013) - OK
XPS 15 9550 (2015) - OK
XPS 15 9560 (2017) - OK
XPS 15 9570 (2018) - OK, so far so good
XPS 15 7590 (2019) - What???
XPS 15 9500 (2020) - What??? Back to the future??? See 2013....
XPS 15 9510 (2021) - OK, back in sequence, I guess....
XPS 15 9520 (2022) - OK, but what are you going to do next year, when you land on 2013's model number?

Ha, ha, ha, ha.....

LoL.
And I don't think they ever tried to make sense of it like amd is trying to do here by explaining the reason behind the code.

AMD ryzen 5 2500u/3500u and Ryzen 7 2700u/3700u only had a difference in number of GPU cores.

Ryzen 7 4700u didn't have SMT.



Compared to that we have a standardised naming scheme, which I think is a good thing. Again, would be nice if there was something that indicated iGPU specs.


As for intel and core i series, people are just too familiar with i5 and i7 without really knowing anything about it due to years of advertising.
They really don't need to make it clear. It would only backfire if they did, especially with P cores and E cores now.
Once upon a time, i7 was 4 cores with HT, i5 was 4 without HT, i3 had 2 with HT.
For u series i7 was dual core with ht. Some i5s had ht too...
When ryzen came along with increased number of cores, intel didn't really have a problem with increasing i7 u to quad cores... meaning they could've done that much earlier, but decided not to because there was no competition...

[ET3D]

What's confusing isn't the naming but the inclusion of different architectures in the same "generation". Given this, the naming is better than any previous naming.

I think that the architecture's digit place is in a logical position. The first digit roughly corresponds to release year, and sure, that's only a marketing digit which helps OEMs sell. Putting the architecture here would have been informative from the customer's perspective, but it would be hell for marketing, as for example Mendocino would be matched to a 2019 family.

The second digit as a performance marker is in a good place, as it's what the buyer wants to know. Architecture affects single-threaded performance, and in that respect a 5600U would be faster than a 5700U, but on the other hand for multi-threading the 5700U is faster than the 5600U. I don't think it would make sense to place the architecture before the performance digit, as then a 4 core with low clocks could be placed before an older architecture 8 core with higher clocks.

So it makes sense to relegate the architecture digit to third. It's better than not putting it there at all, which I think would have been more confusing. So sure, it would be better to have the same architecture for the entire lineup, but if that's not the case, then what AMD has chosen is as good a solution as any.

[Russel]

What's confusing isn't the naming but the inclusion of different architectures in the same "generation". Given this, the naming is better than any previous naming.

I think that the architecture's digit place is in a logical position. The first digit roughly corresponds to release year, and sure, that's only a marketing digit which helps OEMs sell. Putting the architecture here would have been informative from the customer's perspective, but it would be hell for marketing, as for example Mendocino would be matched to a 2019 family.

The second digit as a performance marker is in a good place, as it's what the buyer wants to know. Architecture affects single-threaded performance, and in that respect a 5600U would be faster than a 5700U, but on the other hand for multi-threading the 5700U is faster than the 5600U. I don't think it would make sense to place the architecture before the performance digit, as then a 4 core with low clocks could be placed before an older architecture 8 core with higher clocks.

So it makes sense to relegate the architecture digit to third. It's better than not putting it there at all, which I think would have been more confusing. So sure, it would be better to have the same architecture for the entire lineup, but if that's not the case, then what AMD has chosen is as good a solution as any.

This comment makes the most sense out there xD