Rumor: Ryzen 3000XT chips with higher clock speeds are on the way

midian182

Posts: 9,719   +121
Staff member
Rumor mill: Is AMD about to steal the attention away from Intel’s Comet Lake CPUs by refreshing its Ryzen 3000-series chips? That's the rumor, which claims the Ryzen 9 3900X, Ryzen 7 3800X, and Ryzen 5 3600X are all being revamped with a performance boost.

The rumor comes from a Chinese user posting on the Chiphell forums. While that does mean a heavy dose of salt is required, Twitter user HXL (via Tom’s Hardware) noted that famed overclocked Toppc previously posted the same specs on his Facebook account before deleting the post.

According to the claim, the three ‘Matisse Refresh' processors, which are identified by a “T” moniker, won’t see any core/thread increases as AMD doesn’t want to take sales away from existing Ryzen products. But thanks partly to the continuing refinement of the 7nm FinFET process node, clock speeds have been given a boost.

  • AMD Ryzen 9 3900XT – Base 4.1 GHz / Boost 4.8 GHz (12 Cores / 24 Threads)
  • AMD Ryzen 7 3800XT - Base 4.2 GHz / Boost 4.7 GHz (8 Cores / 16 Threads)
  • AMD Ryzen 5 3600XT – Base 4.0 GHz / Boost 4.7 GHz (6 Cores / 12 Threads)

If the rumor is true, the Ryzen 9 3900XT and Ryzen 7 3800XT come with a 300MHz higher base clock and 200 MHz higher boost than the vanilla versions, while the Ryzen 5 3600XT has a 200 MHz higher base and 300 MHz higher boost.

The processors are also said to support a higher Fabric Clock (FCLK), reaching 2,000 MHz. Overall, the chips offer a 5 to 10 percent increased in performance over their ‘X’ predecessors.

It seems that the XT series will launch at the same MSRP as the current X series, which will see a price reduction. While the new CPUs will compete with Intel’s Comet Lake, it's a little surprising to see a Zen 2 refresh with Zen 3 arriving later this year—unless the latter has been delayed.

Should the rumor prove true, the new processors will be revealed on June 16th, going on general sale on July 7th.

Permalink to story.

 
Why would AMD release the refresh ? Why not ?
It's probably a very low effort and like this article mentions, the 7nm processes continued refinement could indeed enable this. So if they already have enough good dies, why not sell them for a little bit more ?

If the new XT models get roughly the same die quality as the 3950x, the clock speed uplift sounds perfectly reasonable.

From Techspot's own 3950x review:

But when we measured power consumption, that’s when we discovered just how impressive the 3950X is. We’re still blown away by the fact that it offers almost 30% more performance, yet consumes a few watts less than the 3900X. The 3950X really showcases what’s possible with TSMC’s 7nm process.
 
No, Zen 3 hasn't been delayed. This kind of clock boost would require a new stepping (potentially on N7P now, instead of the base N7), so this refresh must have always been part of the plan all along. It's just a way to cheapily & easily smother any potential increased competition + diverted press attention due to Intel's Comet Lake-S launch, until Zen 3 shows up around November.

The level of difficulty & effort for such an endevour is much the same as going from 14nm Polaris 10/20 to 12nm Polaris 30. Aka, the die's floorplan & masks remain untouched, but there's more space between the transistors now, reducing power leakage & allowing power management to be reworked for higher clock-speeds. (Unlike with say Zen+ though, don't expect them to also mess with anything except power management/clocks).

Now what it does suggest, is that Zen 3 prices are likely to be higher than many might expect though. For the first time since Athlon 64 & 64x2, they are going to be absolutely performance dominant in every single workload/metric. With Ryzen 3000 Refresh already more than enough to dispatch Comet Lake-S, expect early Zen 3 parts to have a SERIOUS price premium. AMD's profit margins have been absolute garbage for way too long, and even the improved current situation is simply unsustainable. An utterly unimpeachable halo product like Zen 3 would have the sweet-sweet combination of utter performance dominance (and all the demand that brings) along with the initially high prices & profit margins such performance can fairly demand & that AMD desperately needs.
 
Last edited:
Doesn't the use of XT break AMD's original naming system for the Ryzen line?

2017-03-0218-24-16.jpg

I must admit, I don't recall seeing, for example, a Ryzen 5 1600T and they use the label E on some OEM models, so they've probably dropped the whole thing.
 
On second thoughts, maybe AMD wants to keep the Ryzen 3000 series around as the budget option that Ryzen 2000 is right now ? Not sure how much longer they will be producing on GloFo 12nm. In that sense, a 3300 - 3600 - 3800 and 3900 (with or without XT) line up would make sense even long term if the price is right.

Could be the same thing with the supposed Vega refresh - drop the prices enough and it could become the new Polaris.
 
With the increase in FLCK and High speed DDR4 memory and enhanced clock speeds this should provide a nice little boost.
 
They are getting back better 7nm yields now than before so they are trying to repackage these better-binned chips as a new product to earn some more money. Then they will lower prices on the products with the lower binned 7nm silicon to get rid of this low binned silicon glut. Once the low-binned silicon glut is gone, they will focus on getting rid of the higher-binned silicon to prepare for Zen 3.
 
I suspect Zen 3 will not disappoint in performance at all, but it will disappoint quite some people in pricing. When Zen 3 is out, these better binned Zen 2 CPUs will become AMD's budget option. If AMD can say their budget option is on a par with Intel's Comet Lake, it could be the nail in Intel's coffin.
 
Doesn't the use of XT break AMD's original naming system for the Ryzen line?

View attachment 86661

I must admit, I don't recall seeing, for example, a Ryzen 5 1600T and they use the label E on some OEM models, so they've probably dropped the whole thing.

That graphic was community made. AMD never released an official graphic showing how they name things.
 
Gaming results should get even closer with higher clocks, interested to see the 3600XT and 3800XT.

If AMD can say their budget option is on a par with Intel's Comet Lake, it could be the nail in Intel's coffin.
For non-gamers, which unfortunately, isn't a significant amount of the 'build your PC market'.
They need to be included in more OEM systems, that is where the money's at, business and enterprise.
Another issue there?
An everyday 6-core i5 is more then enough for 90% of business applications, so while Ryzen is great value, AMD is knocking it out of the park but no one's watching the game.
I work for the state of NY and we just did a state wide refresh (well, our refresh team, I just supervised them and assisted, built them a batch file on a USB that import/export all favs, local docs, outlook sig and sticky notes, cause I'm awesome), literally thousands of systems across various sites, from Albany to Syracuse, and they are all HP 440 G6's with i5-9500's/8GB RAM or like mine, i7-9700's with 16GB RAM.
AMD is running out of time.
It's been 3-5 years of Ryzen and most major backbones and enterprise systems are still running on Intel. Most business PC's are still running on Intel.
If Intel does come out with something before AMD can dig the ground up under their feet, it may be too late. This COVID thing is really happening at the worst time for AMD.
 
Last edited:
Intel 10th gen is basically the xt version of 9th gen, but with an even skimpier boost in performance than these.
 
Gaming results should get even closer with higher clocks, interested to see the 3600XT and 3800XT.


For non-gamers, which unfortunately, isn't a significant amount of the 'build your PC market'.
They need to be included in more OEM systems, that is where the money's at, business and enterprise.
Another issue there?
An everyday 6-core i5 is more then enough for 90% of business applications, so while Ryzen is great value, AMD is knocking it out of the park but no one's watching the game.
I work for the state of NY and we just did a state wide refresh (well, our refresh team, I just supervised them and assisted, built them a batch file on a USB that import/export all favs, local docs, outlook sig and sticky notes, cause I'm awesome), literally thousands of systems across various sites, from Albany to Syracuse, and they are all HP 440 G6's with i5-9500's/8GB RAM or like mine, i7-9700's with 16GB RAM.
AMD is running out of time.
It's been 3-5 years of Ryzen and most major backbones and enterprise systems are still running on Intel. Most business PC's are still running on Intel.
If Intel does come out with something before AMD can dig the ground up under their feet, it may be too late. This COVID thing is really happening at the worst time for AMD.
...and most business PC is hacked and locked with ransomware attacks because they used Intel not AMD.
 
Now what it does suggest, is that Zen 3 prices are likely to be higher than many might expect though. For the first time since Athlon 64 & 64x2, they are going to be absolutely performance dominant in every single workload/metric. With Ryzen 3000 Refresh already more than enough to dispatch Comet Lake-S, expect early Zen 3 parts to have a SERIOUS price premium. AMD's profit margins have been absolute garbage for way too long, and even the improved current situation is simply unsustainable. An utterly unimpeachable halo product like Zen 3 would have the sweet-sweet combination of utter performance dominance (and all the demand that brings) along with the initially high prices & profit margins such performance can fairly demand & that AMD desperately needs.
Time may prove your assessment correct; however, I feel inclined to mention the fact that their target with Zen was the enterprise market - aka EPYC as that kind of workload was where Zen initially excelled. IMO, that was a smart business choice for them even though it left gamers feeling ignored. I think that strategy has panned out given AMD's recent HPC wins, not to mention that it is probably the most lucrative part of the computer hardware market.

TR is another story, too, aimed at the "prosumer" and/or workstation class. I have to wonder just how well the TR3990 is selling.

I would like to think that AMD will not pull a sIntel with the lower-end markets and vastly overprice their procs relative to their in-family performance like sIntel did a few years back.

I was around back then for Athlon FX, and yes, they did price it very similarly to sIntel's previous generation of top-end procs; however, it does not mean that they will do that again.

Sometimes, slow and steady wins the race. As I see it AMD is clearly winning the race right now even if it is slow and steady. I just hope the greed bug does not gain a foothold in AMD as it has clearly done in sIntel. If that does happen, I, for one, will be disappointed.
 
Lol with RDNA 2, what 3-4 months away... and with the same naming skew as their GPU lineup, I declare aint happening
 
Lol, this is one of the more silly things I've read this year.
It looks so strange to call me silly for the fact I would prefer the cheaper and secure option, so do not make me list the security bugs Intel processors have, it' s a long list and every people with a little knowledge in computers is aware of it.
 
Back