AMD Announces The Ryzen 3000XT Series

[wizard69]

I really don't understand the point of this release. An extra 100-200 MHz ST boost clock is hardly worth a new SKU, IMO. Unless the all-core boost is significantly higher (but I doubt it is, since base clocks hardly moved).

Maybe they were just trying to come up with some new product to help drive B550 motherboard sales?

Legally they can now market the chip a "NEW".

[artivision]

Why even argue about tubo boost for an unlocked CPU? If you don't want to mess with clocks and voltages, you can just enable PBO. There are countless articles on how to do it.

PBO doesn't clock the Cpu higher. PBO allows more applications to use higher Pstates out of the available clockings by overriding power limits. AutoOC_Max will allow applications like games that use simpler instructions to clock at 4.9Ghz without exceeding 65wp. PBO will override even the 80wp limit of the 65w processors and allow a trash 90-100w consumption without reason. AMD should drop PBO and also drop 100-150mv from Radeons to and free TurboOC.

[wizard69]

Honestly I was expecting a bit more, but I'm not really bothered by the release. It is basically a way to market to gamers as they even suggest water cooling for these chips.

As for the new Zen I still think it i coming this year, it might be late this year but AMD has repeatedly said it is coming. The only reason to back track on that is if the 5nm rumors where real and even AMD has said 7nm so yeah we will see the new architecture this year.

[Paradigm Shifter]

What a mess the processor lineup of AMD had become ...

I think the AMD naming scheme makes rather more sense than the modern Intel naming scheme.

Take the Core i7-1065G7, for example - without looking it up on Ark, all I can tell is that it's a 10-series i7. Or the alphabet soup that the 9-series chips became with the no-letter, E, F, H, HK, K, KF, KFC, KS, T, TE, X, XE suffixes... without a magic decoder ring, how are you supposed to know what each of those are?

I realise that companies need to be able to differentiate between different SKUs internally, and what makes sense to them may seem daft to users, but both Intel and AMD seem to be pretty poor at it from a consumer standpoint. Really I'd like to know the following at a glance: manufacturer, series, generation, model, cores, gpu, hyperthreading. With modern boost systems, the clockspeeds you get will depend heavily on cooler, so they can be relegated to "information elsewhere on the box". A four digit (five, with Intel now 10th gen Core...) number with or without two letter flags would be sufficient to explain most of it.

...

Anyway, slightly more on topic. Good to see that the process has matured enough for AMD to be confident in a move like this. It's much better than what has happened from some manufacturers in the past who have silently changed the SKU without actually telling anyone. I won't be buying one, though, as I've got a 3900X already and the only other Ryzen I might think about buying any time soon is the 65W 3700X to upgrade a box currently running a 1700.

[CochainComplex]

E, F, H, HK, K, KF, KFC, KS, T, TE, X, XE ...

I think Gold is missing?

[novhack]

E, F, H, HK, K, KF, KFC, KS, T, TE, X, XE suffixes...

AMD is not that far with their suffixes count: U, H, HS, G, E, GE, X, XT, WX, AF (unofficial), ...

[NotMine999]

A totally unexciting launch, B550 should have been released last year already. Besides, 4% faster XT CPUs for the original price isn't what I would call exciting either. Wake me up when AM5 arrives.

I was waiting to see some poster say something lame like "I'll just wait for...." but this post by ms178 will do.

[Paradigm Shifter]

AMD is not that far with their suffixes count: U, H, HS, G, E, GE, X, XT, WX, AF (unofficial), ...

Not sure the AF really counts, since that seems to be more of a stepping identifier? (Like B3, G0 was for the legendary Q6600 back in the days of Core 2)

...but both Intel and AMD seem to be pretty poor at it from a consumer standpoint...

I wasn't saying AMD were innocent. Both sides really need to clean up their naming schemes for consumers.

I remember when AMD release the Radeon 5000 series (that induced EyeFinity to the world, and nVidia scrambled around to bodge "Surround" together across SLI... the six (mini)-DisplayPort 5870 card could have been called the "5876" which would have actually made some sense. They went with "5870 EyeFinity 6 Edition", which was a bit of a mouthful.

To be fair, if we strip out mobile and "HEDT" grade CPUs, AMD fairs somewhat better with just no-letter, X and XT... while Intel still has no-letter, F, K, KF, KFC, KS, and T. Ah, but I forgot the APUs, with that, so valid point. AMD still has G and GE as well.

Anyway, both companies are somewhat lacking at sensible, consumer friendly naming schemes.

edit: what are the naming schemes for Epyc? Just four digits I think? Intel has Bronze, Silver, Gold, Platinum on top of the four digit identifier. But I don't usually worry too much about Xeon chips, 'cause I can't afford them.

[nomadewolf]

This is literally what intel has been doing for half a decade now. Hear ye, hear ye! Announcing our NewLake processors! They are the exact same uarch as OldLake, same process node as OldLake, same gfx as OldLake, but clock is 100 Mhz faster! Innovation!! Upgrade now for only $499!!

Except Intel would call it the 4000X series instead of just adding a T...
AMD is being frontal: minor name change for minor actual change, and the MSRP is the same...
If you want it, good. If not, just pick up the older parts that are cheaper.

[novhack]

Not sure the AF really counts, since that seems to be more of a stepping identifier? (Like B3, G0 was for the legendary Q6600 back in the days of Core 2)

AF is not an official suffix by AMD. They refreshed some first gen Ryzen processors on a 12nm node. If I remember correctly it's basically second gen chips but with lower clocks.
They didn't differentiate these "new" processors in any way so community created AF suffix and even some e-shop accepted it and sell them with that name.