Discussion Speculation: Zen 4 (EPYC 4 "Genoa", Ryzen 7000, etc.)

[Vattila]

Except for the details about the improvements in the microarchitecture, we now know pretty well what to expect with Zen 3.

The leaked presentation by AMD Senior Manager Martin Hilgeman shows that EPYC 3 "Milan" will, as promised and expected, reuse the current platform (SP3), and the system architecture and packaging looks to be the same, with the same 9-die chiplet design and the same maximum core and thread-count (no SMT-4, contrary to rumour). The biggest change revealed so far is the enlargement of the compute complex from 4 cores to 8 cores, all sharing a larger L3 cache ("32+ MB", likely to double to 64 MB, I think).

Hilgeman's slides did also show that EPYC 4 "Genoa" is in the definition phase (or was at the time of the presentation in September, at least), and will come with a new platform (SP5), with new memory support (likely DDR5).

What else do you think we will see with Zen 4? PCI-Express 5 support? Increased core-count? 4-way SMT? New packaging (interposer, 2.5D, 3D)? Integrated memory on package (HBM)?

Vote in the poll and share your thoughts!

 

[Timorous]

Let me put it this way. If AMD was seeing an easy +30% today, then why say half that? It contradicts pretty much their entire history of marketing. And as I already pointed out, if you believe higher clock speeds are possible, that alone makes him wrong.

What do you think would happen to 5000 series and 5800X3D sales if AMD said zen4 performance was 30% greater?

 

[eek2121]

What do you think would happen to 5000 series and 5800X3D sales if AMD said zen4 performance was 30% greater?

They would still sell every chip they make.

The reasons they haven't provided details are:

1. Specifications aren't finalized.
2. Keep quiet to maintain competitive advantage.

 

[gdansk]

Dr. Lisa Su(not known to be a Liar) has twice mentioned ALL CORE 5Ghz+ while gaming(at CES 2022 8 core prototype and recently 16 Core Prototype.

Also the Blender Render would have need to be at 5 ghz All core for the 16C/32T to be 46% faster than a 12900K

Why do you assume 5GHz+ all cores means 5.5GHz all core? 5020MHz all core would match her statement too.

They reduced the gap between lightly loaded and entirely loaded with the 88w higher power limit. But they'll still have lower all core clocks when at full tilt than a lightly threaded workload (which is what they showed running at 5520)

 

[eek2121]

Apparently, Robert also stated multicore improvements of 40+%. That is pretty juicy.

 

[Exist50]

Because it's a good idea to Market a yet to be released CPU months in advanced to give your competitor time to do adjustments? Yeah Right..

Apparently AMD thought so, because that's precisely what they did. I didn't know the denial could run so deep as to pretend the whole event didn't happen.

 

[nicalandia]

Why do you assume 5GHz+ all cores means 5.5GHz all core?

I just generalized/Rounded the speed of two Prototypes. The early one done on January was a 8 Core prototype said to be boosting at 5 Ghz All Cores and Then the 16C Pre-Production Prototype was seen boosting at 5.5Ghz(was later confirmed to be all cores)

Also

 

[Exist50]

Then the 16C Pre-Production Prototype was seen boosting at 5.5Ghz(was later confirmed to be all cores)

They said all game threads were between 5.2-5.5GHz, and as previously established, that game doesn't use 8 cores, much less 16.

 

[nicalandia]

Apparently AMD thought so, because that's precisely what they did. I didn't know the denial could run so deep as to pretend the whole event didn't happen.

Are we really going to debate on "Marketing" Semantics? I call it a Sneak Peak.

 

[SmokSmog]

Yes all cores were boosting to 5,5ghz, one by one not all at once 😉

 

[szrpx]

AMD Corporation Quartz - Geekbench

Benchmark results for an AMD Corporation Quartz with an AMD Eng Sample: 100-000000479-13 processor.

browser.geekbench.com

I know this is probably old news, but has anyone debunked the actual clockspeed of this test run? Or how it compares to Milan / Zen 3 at the same frequency?

 

[nicalandia]

AMD Corporation Quartz - Geekbench

Benchmark results for an AMD Corporation Quartz with an AMD Eng Sample: 100-000000479-13 processor.

browser.geekbench.com

I know this is probably old news, but has anyone debunked the actual clockspeed of this test run? Or how it compares to Milan / Zen 3 at the same frequency?

Seems Legit for a Zen4 Genoa with 32 cores at low clocks

 

[Doug S]

They would still sell every chip they make.

The reasons they haven't provided details are:

1. Specifications aren't finalized.
2. Keep quiet to maintain competitive advantage.

#1 makes perfect sense, it might be a 16% gain today but they think with another stepping to clean up a few issues it will be higher when released, but they don't want to commit to anything and get egg on their face so they are being conservative until they know for sure.

#2 makes no sense at all. AMD is selling every Zen 3 they make, and will continue selling them even when Zen 4 is on the market, so they have no worries about being Osbourned. If they want to steal share from Intel they should want to advertise the best possible numbers they can, because they would want people considering buying Intel today to decide they need to wait for Zen 4 before making their decision.

There's no world where underselling performance "maintains competitive advantage". It only helps if you want to deliberately hamstring performance in a released product, like Intel has been accused of doing in the past when they effectively had no competition.

 

[Abwx]

#1 makes perfect sense, it might be a 16% gain today but they think with another stepping to clean up a few issues it will be higher when released, but they don't want to commit to anything and get egg on their face so they are being conservative until they know for sure.

Personaly if that were actually 16% i wouldnt state +15%, the margin is just too low to be a guaranted figure, better to avoid any litigation...

 

[biostud]

You quote AMD's Director of Technical Marketing, commenting after a marketing presentation, as proof they aren't doing marketing? Are you just trying to mess with us?

Well, I'm pretty sure that not a single one of the participants in this thread would buy a new computer without having done some research.

Sure, I'm pretty certain that I'll buy a zen4 setup, but should it for some reason be a Bulldozer and Raptor Lake surpassing it by 30%, then Raptor lake it is.

My point being if they didn't release anything at computex, we would be disappointed, if they release numbers they can't deliver at launch, we will be disappointed at launch, so what they can release are small tidbits of underwhelming information. So damned if you do and damned if you don't.

 

[nicalandia]

My point being if they didn't release anything at computex, we would be disappointed, if they release numbers they can't deliver at launch, we will be disappointed at launch, so what they can release are small tidbits of underwhelming information. So damned if you do and damned if you don't.

I agree

 

[Doug S]

Personaly if that were actually 16% i wouldnt state +15%, the margin is just too low to be a guaranted figure, better to avoid any litigation...

Litigation is never a concern, all they have to do is get 16% in whatever benchmark(s) they used which they don't even have to disclose. We've all seen inflated claims from many vendors which are only true in some circumstances, there were never any lawsuits.

Who would even have standing to sue? It was a marketing event, those CPUs aren't for sale yet so no one can claim they were defrauded. The bar for investors suing is very high, if every company that makes puffed up claims could be sued 90% of the population would have to be lawyers and judges to handle all the cases.

 

[Abwx]

Litigation is never a concern, all they have to do is get 16% in whatever benchmark(s) they used which they don't even have to disclose. We've all seen inflated claims from many vendors which are only true in some circumstances, there were never any lawsuits.

Who would even have standing to sue? It was a marketing event, those CPUs aren't for sale yet so no one can claim they were defrauded. The bar for investors suing is very high, if every company that makes puffed up claims could be sued 90% of the population would have to be lawyers and judges to handle all the cases.

If they say +15% then it must be more than 15% for all SKUs counted as Ryzen 7000, even the 6C, and all chips wont clock as high as 5.5GHz.

If we take this as the top SKU max boost then expect the lower parts to be at 5GHz, same way as Zen 3 wich has 300-400Mhz lower boost frequency for the 6C.

 

[inf64]

It's kind of pointless to go back-and-forth with the number AMD provided as we simply don't know. It's >15% so it could be 16 or it could be 25%+, we simply don't know.
It's ok to not believe it will be higher than some specific number but to claim it cannot be higher for whatever reason is pointless.

 

[JoeRambo]

AMD cannot afford to have low clocks anywhere, the Skylake days are over and in it's place is brutal GC core backed by efficiency cores.
The reality is that Intel will fight 7600X with 13600K that is rumored 6P + 8E. 7800x will be assaulted by 13700K with 8P + 8E config and so on, increase of E cores trickles down the line.

7950x has 8vs8 (vs 8P) that leaves 8 Z4 cores fighting 16E
7900x has 6+6 (vs 8P) that leaves 4 cores fighting 8E
7800x has 8 (vs 8P) that leaves 8E cores
7600x has 6 (vs 6P)that leaves 8E cores

It is fine to discuss CB scores of 35-40k and high end SKUs, but bulk of sales is not there. It's the lower end SKUs that need clocks the most to not loose throughput tests in disastrous way. AMD is really walking on thin ice here, their ST advantage has evaporated and seems that it is not really coming back and they need to push clocks and TDP to what is most likely stupid efficiency region.
On top of that new platform with DDR5 that makes selling anything with less than 8C an uphill battle. The SKUs and pricing are going to be very interesting this timem we might get 7700x real quick this time. Viva the competition!

 

[inf64]

Here is Robert Hallock stating that they are getting 40+% higher performance right now ( with the current state of silicon) : LINK
I guess he is referring to Zen4 vs Zen3 in MT workloads but that's really encouraging as it implies that multi core boost is much higher (north of 20% vs 4.2Ghz on 5950X) and IPC should be at minimum >10%

 

[nicalandia]

It's kind of pointless to go back-and-forth with the number AMD provided as we simply don't know. It's >15% so it could be 16 or it could be 25%+, we simply don't know.
It's ok to not believe it will be higher than some specific number but to claim it cannot be higher for whatever reason is pointless.

Would you stop making sense and get with the program?

 

[PJVol]

I know this is probably old news, but has anyone debunked the actual clockspeed of this test run?

What do you mean actual? Geekbench browser reported 3430Mhz.

Anyways, it seems AMD managed to squeeze more from the core design regarding energy efficiency.
With Zen2 they did achieve +15% IPC and at the same time reduced Cac by 9%. Curious about current state of design optimizations and improvements, looking at these 45+% MT numbers.

 

[szrpx]

What do you mean actual? Geekbench browser reported 3430Mhz.

Ah, I see. Thanks for clearing that up.

 

[biostud]

we might get 7700x real quick this time. Viva the competition!

And 7800X3D?

 

[DrMrLordX]

brutal GC core

. . . lol