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

Page 228 - Seeking answers? Join the AnandTech community: where nearly half-a-million members share solutions and discuss the latest tech.

Vattila

Senior member
Oct 22, 2004
799
1,351
136
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).

Untitled2.png


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! :)
 
Last edited:
  • Like
Reactions: richardllewis_01

Anhiel

Member
May 12, 2022
63
17
41
I just wish there would be more detailed tests around with for example 5600X with clockspeed increments and power consumption all the way up to stock limits and separately OC limits.
Also having better data on all core clock speed for stock and OC in general for all SKUs would be great.
So far 99% of tests out there don't provide the one I need.
 
Last edited:

eek2121

Platinum Member
Aug 2, 2005
2,904
3,906
136
Netburst says hi.

AMD might be beating Intel at the frequency game for the first time in a long time. 5.52 ghz on more than 1 core is higher than even the 12900ks.

At any rate, based on comments made by AMD, it is clear why they did not provide specific details in terms of performance. They don't know where they are going to pin frequencies as of yet. They will likely need a larger sample of chips in order to determine cut-off. If they are showing 5.52 GHz now, it is very likely that is a number they are comfortable they can hit. It is likely the final version may boost even higher.

I'm interested in other improvements in terms of power consumption (idle power, power under load, IF power consumption, etc.)
 
  • Like
Reactions: lightmanek

Exist50

Platinum Member
Aug 18, 2016
2,445
3,043
136
dude if you really must go down the road of this sacred leakerbusting, at least try to be bothered to admit, that the 37% figure was mentioned as an 'as high as' metric, so at the high end of what's possible, not as a general 37% increase in ST workload performance.
Come on, man. He's really not worth the effort you're putting into defending him. If his minimum IPC number was accurate, then given the clocks shown already, AMD could probably claim +30% performance. And if the clocks actually end up higher, then it'll be above the top of his range! He was wildly off base, and it's not the first time. So let's just accept that and move on.

You might wanna re-watch his video about that.
Lol, then I better be quick, or he'll delete it before I do.
 
  • Like
Reactions: mikk and JoeRambo

Mopetar

Diamond Member
Jan 31, 2011
7,797
5,899
136
I have no need for more than 8 cores, so I hope the 7800X will clock just as high as 7950X.

It seems likely that it will clock similarly well. Obviously AMD will want to use the best silicon for the top chips to upsell customers, but if the 7950X (or whatever it is called) can hit 5.5 GHz, then the 7800X should hit at least 5.4 GHz based on past products.
 

deasd

Senior member
Dec 31, 2013
513
724
136
Come on, man. He's really not worth the effort you're putting into defending him. If his minimum IPC number was accurate, then given the clocks shown already, AMD could probably claim +30% performance. And if the clocks actually end up higher, then it'll be above the top of his range! He was wildly off base, and it's not the first time. So let's just accept that and move on.

Lol, then I better be quick, or he'll delete it before I do.

Please don't blame MLID.....because just now I unintentionally find a even much more exaggerate Zen4 IPC number than MLID's, which was already being reported at least 15 months ago:



A Zen 4 Genoa engineering sample reportedly performed 29% faster than an existing Zen 3 CPUs at the same clock speeds and core counts.

but the original source/leak was erased(or a fake?). Maybe MLID's number should be called conservative now even if he's wrong.

edit: well yeah it's from chipsandcheese

however I was told from a trusted source that a Genoa engineering sample (Zen 4 server chip) was 29% faster than a Milan (Zen 3) chip with the same core config at the same clocks.

busted!
 
Last edited:

Exist50

Platinum Member
Aug 18, 2016
2,445
3,043
136
Please don't blame MLID.....because just now I unintentionally find a even much more exaggerate Zen4 IPC number than MLID's, which was already being reported at least 15 months ago:





but the original source/leak was erased(or a fake?). Maybe MLID's number should be called conservative now even if he's wrong.

edit: well yeah it's from chipsandcheese



busted!
That even more egregiously wrong "leaks" exist does not make his any closer to reality. This is well beyond any reasonable margin for error.
 

maddie

Diamond Member
Jul 18, 2010
4,723
4,628
136
Please don't blame MLID.....because just now I unintentionally find a even much more exaggerate Zen4 IPC number than MLID's, which was already being reported at least 15 months ago:





but the original source/leak was erased(or a fake?). Maybe MLID's number should be called conservative now even if he's wrong.

edit: well yeah it's from chipsandcheese



busted!
Concerning the chips&cheese statement.

Without having much context, this could be true. Extra cache engineering sample?
 
  • Like
Reactions: lightmanek

SmokSmog

Member
Oct 2, 2020
56
88
61
So much hopium here, 6ghz single core 5,5ghz+ all core oc LOL.

We will end up with 16 core Zen4 boosting to 5,5ghz for split second with precision boost 1,5V under light load and 5-5,2ghz all core OC on water cooling.
 
  • Like
Reactions: Hulk

nicalandia

Diamond Member
Jan 10, 2019
3,330
5,281
136
So much hopium here, 6ghz single core 5,5ghz+ all core oc LOL.

We will end up with 16 core Zen4 boosting to 5,5ghz for split second with precision boost 1,5V under light load and 5-5,2ghz all core OC on water cooling.

No need to OC for the 5.5 Ghz All core boost. It was actually an Easy Achievement for Pre-Production Prototype CPUs.
 

gdansk

Golden Member
Feb 8, 2011
1,991
2,359
136
No need to OC for the 5.5 Ghz All core boost. It was actually an Easy Achievement for Pre-Production Prototype CPUs.
That's not what they said. It was easy to hit 5.5GHz in a few threads. The threads running the game. It's probably only a few.

And it is Robert, who already had to do a correction about his confirmation of 170W PPT.
I wouldn't trust anything vague like 'easy' when coming out of his mouth.
 
  • Like
Reactions: SmokSmog

Exist50

Platinum Member
Aug 18, 2016
2,445
3,043
136
And you know this because.....? Yes, because AMD was Sandbagging using a >15% in ST on CineBench R23 Right???
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.
 

eek2121

Platinum Member
Aug 2, 2005
2,904
3,906
136
That's not what they said. It was easy to hit 5.5GHz in a few threads. The threads running the game. It's probably only a few.

And it is Robert, who already had to do a correction about his confirmation of 170W PPT.
I wouldn't trust anything vague like 'easy' when coming out of his mouth.

It maintained speeds of around 5.2-5.5 ghz on multiple cores for an extended period of time. Looked easy to me. Preproduction silicon even.

Not even the 12900ks can do that without being overclocked. The 12900ks MIGHT hit 5.5ghz on 1 core if you are lucky, and it isn’t going to do that for any real workload.
 
  • Like
Reactions: lightmanek

SmokSmog

Member
Oct 2, 2020
56
88
61
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.

4f5298993ab902699439eaae9ab543444b2bf9d892199efa08508cf246ce6ba0.jpg

No need to OC for the 5.5 Ghz All core boost. It was actually an Easy Achievement for Pre-Production Prototype CPUs.
4,9-5,05ghz is easy to achieve on stock boost for 5950X, can you all core OC it to that ?
 
Last edited:

gdansk

Golden Member
Feb 8, 2011
1,991
2,359
136
It maintained speeds of around 5.2-5.5 ghz on multiple cores for an extended period of time. Looked easy to me. Preproduction silicon even.

Not even the 12900ks can do that without being overclocked. The 12900ks MIGHT hit 5.5ghz on 1 core if you are lucky, and it isn’t going to do that for any real workload.
With only a few cores loaded even the 5950X clocks pretty high.
I very much doubt the 16 core Zen 4 part, even with 230W, will be 5.5GHz all core without PBO/OC.
 
  • Like
Reactions: lightmanek

Hans Gruber

Platinum Member
Dec 23, 2006
2,092
1,065
136
AMD might be beating Intel at the frequency game for the first time in a long time. 5.52 ghz on more than 1 core is higher than even the 12900ks.

At any rate, based on comments made by AMD, it is clear why they did not provide specific details in terms of performance. They don't know where they are going to pin frequencies as of yet. They will likely need a larger sample of chips in order to determine cut-off. If they are showing 5.52 GHz now, it is very likely that is a number they are comfortable they can hit. It is likely the final version may boost even higher.

I'm interested in other improvements in terms of power consumption (idle power, power under load, IF power consumption, etc.)
Word on the street is that the Intel Raptor Lake will do 5.7-5.8ghz on a single core. The real competition will be all core max speed for both AMD and Intel.
 
  • Like
Reactions: lightmanek

nicalandia

Diamond Member
Jan 10, 2019
3,330
5,281
136
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.
News Flash. They ARE NOT Trying to Market this CPU as of Yet.

"We were deliberately conservative with our number on single-thread performance. We do intend to publish the exact breakdown of IPC vs. frequency contribution later in the summer, " Robert Hallock
 
  • Like
Reactions: lightmanek

Exist50

Platinum Member
Aug 18, 2016
2,445
3,043
136
News Flash. They ARE NOT Trying to Market this CPU as of Yet.

"We were deliberately conservative with our number on single-thread performance. We do intend to publish the exact breakdown of IPC vs. frequency contribution later in the summer, " Robert Hallock
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?
 

nicalandia

Diamond Member
Jan 10, 2019
3,330
5,281
136
That's not what they said. It was easy to hit 5.5GHz in a few threads. The threads running the game. It's probably only a few.
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
 

nicalandia

Diamond Member
Jan 10, 2019
3,330
5,281
136
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?
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..

You are really beyond Hope(and not just dumb) to believe AMD will be telling You and anybody else what is their Real Performance this early in the game.
 

eek2121

Platinum Member
Aug 2, 2005
2,904
3,906
136
Word on the street is that the Intel Raptor Lake will do 5.7-5.8ghz on a single core. The real competition will be all core max speed for both AMD and Intel.

I will believe it when I see it. I heard similar claims about Alder Lake.

They will need every single hertz possible in order to catch up to Zen 4.
 
  • Like
Reactions: lightmanek