Speculation: Ryzen 4000 series/Zen 3

[Hitman928]

I know but nervertheless a desktop system will be a bit faster. Sometimes not even because of better RAM latencies, sometimes because of small clock speed drops. At 28W i7-1185G7 still has a few clock speed drops in Cinebench 1T:

Intel Core i7-1185G Tiger Lake im Test: Testergebnisse und Benchmarks

Tiger Lake im Test: Testergebnisse und Benchmarks / Single-Core-Leistung: ein neuer Rekord / Multi-Core-Leistung: AMD bleibt vorne

www.computerbase.de

Zen2 struggles to maintain its max boost speed in CB single thread as well. We'll have to wait for reviews to get the full scoop.

 

[cytg111]

Better than tigerlake around the block AND as many cores as your heart desires.
Its gonna be a another bad year for Intel.

 

[inf64]

It's a safe bet to assume 5% better IPC than Tigerlake at this point. I guess AVX512 stuff will be last area where Icelake derivatives will be ahead. Unless Tigerlake outclocks Cezzane by more than 5% while running at lower power and adding 2x cores and Rocektlake (which will have lower IPC than Tigerlake) magically clocks to 5.5Ghz, AMD will have performance and perf./watt lead in desktop, server and laptop segments.

 

[mikk]

It's a safe bet to assume 5% better IPC than Tigerlake at this point. I guess AVX512 stuff will be last area where Icelake derivatives will be ahead. Unless Tigerlake outclocks Cezzane by more than 5% while running at lower power and adding 2x cores and Rocektlake (which will have lower IPC than Tigerlake) magically clocks to 5.5Ghz, AMD will have performance and perf./watt lead in desktop, server and laptop segments.

I'm not sure about the performance lead in gaming because the gap is really small compared with 10900k on AMDs marketing slide. If it isn't caused by a GPU bottleneck RKL-S might beat Zen 3, especially if the rumoured 5.3 Ghz turns out to be true.

 

[nicalandia]

People complaining about low clock speed. to get 640 ST Cinebench the 3950X would need to be OC to 5.6 Ghz…. Guys speed is not all that matters but IPC

 

[inf64]

I'm not sure about the performance lead in gaming because the gap is really small compared with 10900k on AMDs marketing slide. If it isn't caused by a GPU bottleneck RKL-S might beat Zen 3, especially if the rumoured 5.3 Ghz turns out to be true.

The lead is small in a few tests as they are GPU bound even at 1080p. The more CPU bound tests produce higher difference as you can see. I think that rocket lake will stand no chance, it's rumored by Chinese leakers to be only around ~10% better IPC wise than skylake ( most likely apps and not games). Even at 5.5Ghz Cypress Cove( Rocketlake) will have a mountain to climb as Zen3 is better than a core such as Willow Cove.

 

[yeshua]

I'm not sure about the performance lead in gaming because the gap is really small compared with 10900k on AMDs marketing slide. If it isn't caused by a GPU bottleneck RKL-S might beat Zen 3, especially if the rumoured 5.3 Ghz turns out to be true.

Rocket Lake will definitely beat Ryzen 5000 in raw performance at the cost of a much higher power consumption now that it's rumored to still be using the 14nm ++++ node.

 

[Markfw]

Rocket Lake will definitely beat Ryzen 5000 in raw performance at the cost of a much higher power consumption now that it's rumored to still be using the 14nm ++++ node.

I wouldn't bet on that....

 

[dnavas]

And big surprise, NO SMT4 ! (not really)

Although to be fair the presentation lacked a lot in technical detail, so I expect we'll continue to hear from the peanut gallery. "Wider FE" tell me more? Marketing took front seat, which explains the pricing as well, and is exactly as I feared -- the piper is being paid. The downside is exactly as yeshua put it -- "Customers first, my buttocks." AMD can't afford to lose customer enthusiasm, and raising prices in challenging times is tone deaf.

But, that's a problem for down the road when they have competition again. I'm far more interested in the details which we've heard nothing about, and presumably won't for another month. :gah: To continue to be relevant, AMD needs to leverage the developer market. Bandwidth is rising faster than computation speeds, and the market is going to need a "1000 points of light" [well, ok, dozens] set of solutions for managing streaming loads. This kind of technically lightweight presentation is going to get harder to pull off. At least I hope it will -- certainly the console integration will pay off wrt fs reads, they just need the same to happen for networking (Mellanox getting swallowed won't make that easier for them), and they need a far better way to leverage and integrate gpu-style vector units.

Then again, that's all for server platforms anyway, so maybe I'm just never going to see the kind of presentation I'm interested in. Maybe the future will be all about laptop battery life, idle performance, and gameplay 8k leapfrogging. I suppose that'll be fine for the horserace crowd, but I can't imagine a duller future :| I presume ymmv!

 

[nicalandia]

The lead is small in a few tests as they are GPU bound even at 1080p. The more CPU bound tests produce higher difference as you can see. I think that rocket lake will stand no chance, it's rumored by Chinese leakers to be only around ~10% better IPC wise than skylake ( most likely apps and not games). Even at 5.5Ghz Cypress Cove( Rocketlake) will have a mountain to climb as Zen3 is better than a core such as Willow Cove.

Rocket Lake still on 14nm and Zen 3> Willow Cove> Cypress Cove

 

[nicalandia]

AMD can't afford to lose customer enthusiasm, and raising prices in challenging times is tone deaf.

People All Up in Arms for a $50 increment for a 20% IPC Gain.. Yeah let's forget that Intel used to charge $1500 for an 10 Core Processor not so long ago

 

[Hitman928]

Although to be fair the presentation lacked a lot in technical detail, so I expect we'll continue to hear from the peanut gallery. "Wider FE" tell me more? Marketing took front seat, which explains the pricing as well, and is exactly as I feared -- the piper is being paid. The downside is exactly as yeshua put it -- "Customers first, my buttocks." AMD can't afford to lose customer enthusiasm, and raising prices in challenging times is tone deaf.

I don't know about other countries, but in the U.S. people have actually been on a spending spree in terms of home equipment (electronics, exercise, home theater, etc.). No matter how people may feel about the prices, the market this year has actually seen more demand for these products and even at the high end of the price spectrum people have been buying like crazy. Whether you feel it is right or wrong, the market doesn't seem to flinch at increased prices this year.

 

[yeshua]

It's funny AMD has leaked yet to be announced 5900XT and 5950XT in their own presentation:

 

[Kedas]

If consoles are on Zen2 they are missing out on a good zen3 upgrade, although 4K TV gaming won't make much of a difference....

 

[inf64]

It's funny AMD has leaked yet to be announced 5900XT and 5950XT in their own presentation:

Those could be +100Mhz higher base/boost clocked models. Of course with 100 USD higher prices

 

[dnavas]

People All Up in Arms for a $50 increment for a 20% IPC Gain.. Yeah let's forget that Intel used to charge $1500 for an 10 Core Processor not so long ago

Really, "All Up in Arms" -- uppercase, even? I think you're over-reaching. Did you not read the "I'm far more interested" part of that post? There's a lot more there there. If I'm up in arms about something, it's about the utter lack of technical detail. Regardless I'm not "All Up in Arms". Goodness -- sensitive much?

But, if you want to talk pricing ... I don't know how old you are, but I'm from a time when processors got 100% faster every year and a half, and prices were as likely to slip as rise. So, yeah, 20% is not impressive, neither is paying twice as much for twice the cores. The whole point of the industry is perf/$ -- 20% perf gain minus 10% price rise == 10% value increase. As I posted somewhat long ago, my only real hope is that 39x0 TR pricing will see a reset. We'll see :shrug:

 

[Hitman928]

It's funny AMD has leaked yet to be announced 5900XT and 5950XT in their own presentation:

Pretty sure that's just a typo.

 

[nicalandia]

Yeah, 20% is not impressive, neither is paying twice as much for twice the cores.

Really? You Intel Fans have a short memory span or are you as thick as you sound? How about paying $20,000 for a 56 Core Xeon Processor? How about that. I am sure guys like you complained that the 64 Core Rome Based EPYCs were $6000 a pop. How about 64C Thread Ripper being $4000? Yeah keep complaining about $50 price hike.

 

[dnavas]

I don't know about other countries, but in the U.S. people have actually been on a spending spree in

That's a really excellent point. You're right. I'd call consumers (not to mention the stock market) out and name this for what it is, but folks around here seem sensitive, so maybe I'll Obama-ize it and just say that present-day consumers appear to discount the circumstances in which they find themselves and behave in a manner that calls into question the degree of self-reflection in which they are engaged; that companies find themselves in a position to take advantage and choose to do so is hardly reflective solely on them.

That said, there are some gaming increases that rise well above the 20% improvement bar, so while that improvement is not likely to benefit my workloads, I do think there are rational reasons to pay decent money for an upgrade. Perf per watt sees a nice rise, and we (again, not to beat a dead horse) have yet to hear anything technical about, say, how multicore clocks behave.

Really? You Intel Fans [rant]

 

[NTMBK]

From the Anandtech article:

There is also a third factor, in that the number 4 in China is often seen as an unlucky number due to the sound of the number being close to the sound of the Chinese word for ‘death’. The Chinese market has been a key growth market for AMD, and so aligning its latest generation hardware out of the ‘4000’ category will remove a potential negative to sales. This is also perhaps why so few of AMD’s processors contain the number 4.

Well that explains the SMT4! Obviously it's in the chip, but AMD had to disable it so that they don't spook China

 

[teejee]

Based on the Cinebench 1T score, Zen 3 has a 5% higher IPC than Willow Cove at 4.8 Ghz....in Cinebench at least. The game comparison with 10900k isn't clear....I wonder if it was bottlenecked by the GPU, beside Counterstrike and League of Legends the improvement over 10900K is really small. With DDR4-3200 10900K would easily match or beat this apart from Counterstrike and League of Legends. This is still a concern.

They used DDR4-3600 in the test for both systems.
This was not an Intel presentation...

 

[therealmongo]

Let me be brutally honest. AMD is no different than Intel in terms of dictating prices when they have the performance crown.

The pricing for the Ryzen 5000 series:

5600X: $300 - very close to the price of the 3700X which featured two more cores.
5800X: $450
5900X: $550
5950X: $800

All priced $50 higher than their Ryzen 3000 counterparts. What's more, there's no sign of 5700X which was a sweet spot for the previous gen Ryzen CPUs. Either you pay $50 more for the 3600X alternative or you have to pay whopping $120 more to get just two more cores.

Customers first, my buttocks. More like profits first now that Intel still cannot solve their 10nm node.

Really, just really.

The World does not revolve around the consumer, its a two edge sword, producer and consumer. AMD clearly have the best CPU on the market, if it was Intel it wouldnt be a $50 increase in price but a $200 increase in price so quit the crying.

And big surprise, NO SMT4 ! (not really)

Hahaahhhhaaahahaa, you could'nt help yourself, cant blame you

Rocket Lake will definitely beat Ryzen 5000 in raw performance at the cost of a much higher power consumption now that it's rumored to still be using the 14nm ++++ node.

Please can I have the lottery numbers?

I have no issue you saying that you believe that Rocket Lake will "beat" the Ryzen 5000 in raw performance, but to say "definately", who are you yeshstradamus ?

 

[mikk]

Rocket Lake will definitely beat Ryzen 5000 in raw performance at the cost of a much higher power consumption now that it's rumored to still be using the 14nm ++++ node.

RKL-S might beat Zen 3 in gaming but overall this chip won't be competitive. They need ADL-S as fast as possible.

 

[dzoni2k2]

Really, "All Up in Arms" -- uppercase, even? I think you're over-reaching. Did you not read the "I'm far more interested" part of that post? There's a lot more there there. If I'm up in arms about something, it's about the utter lack of technical detail. Regardless I'm not "All Up in Arms". Goodness -- sensitive much?

But, if you want to talk pricing ... I don't know how old you are, but I'm from a time when processors got 100% faster every year and a half, and prices were as likely to slip as rise. So, yeah, 20% is not impressive, neither is paying twice as much for twice the cores. The whole point of the industry is perf/$ -- 20% perf gain minus 10% price rise == 10% value increase. As I posted somewhat long ago, my only real hope is that 39x0 TR pricing will see a reset. We'll see :shrug:

What time was that? I'm not exactly young myself and I distinctly remember pretty pathetic performance upgrades in the past. What you usually got was a 100MHz clock bump and 0 IPC increase.

 

[dnavas]

What you usually got was a 100MHz clock bump and 0 IPC increase.

Off topic, but, I wrote code on machines with turbo buttons on the front panel. That was long before 100Mhz was even a thing! Now give me back my cane and get off my lawn!