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!

 

[tamz_msc]

How big is core+L2 for the Apple M2?

 

[inf64]

I must quote our friend reb0rn here:

@inf64
graph made by AMD fanboy as many here still promote and FUD intel for their own reason, there is <1% chance 16/32 7950x is faster then 24/32 13900k

As can be seen, you sir have failed. Better luck next time, and I suggest that you use your brain .

 

[nicalandia]

How big is core+L2 for the Apple M2?

About 4 mm2

 

[Vope45]

More importantly, ~35% overall performance (as indicated in the slide) over Zen 3 @170w @Abwx

Also:
- ~15% average gaming performance over Zen 3 (which other chip is 15% average over Zen 3? 5800x3d, which happens to trade blows with ADL + DDR5). I'm optimistic for RPL. No wonder the X3D is looking to make an almost immediate showing this time around.
- 170w TDP/230w PPT (to allow Zen 4 chips to leg room to do their thing). So, the chip power consumption debate should end tonight.

I think 15% over 5950x, over the 5800x3d might be too much.

 

[Abwx]

More importantly, ~35% overall performance (as indicated in the slide) over Zen 3 @170w @Abwx

That s 45% over the 5950X in Cinebench R23 since the 35% was related to this load, and 41% over the 12900K.

 

[Vope45]

That was in the anandtech liveblog. The actual quote, at -12:10 from the end, was "for 2024". Sorry about that.

Huge difference for sure. Another birdie told me march 2024

 

[Carfax83]

RAM doesnt matter much in Cinebench, here that s 41% and 45% faster than 12900K and 5950X respectively.

I'm not talking about Cinebench. I'm talking about the gaming benchmarks. Memory speed can make a substantial difference in results and the 12900K performance looks depressed.

That said, I can't deny that the Zen 4 presentation looked EXTREMELY impressive. Seems like a well balanced architecture with high clock speeds, high ipc and great performance per watt.

Of course I will wait for reviews, but so far it looks promising.

 

[Exist50]

Huge difference for sure. Another birdie told me march 2024

AMD's been on a year and a half cadence, so that would make sense for them to target, if nothing else.

 

[Vope45]

AMD's been on a year and a half cadence, so that would make sense for them to target, if nothing else.

Same little birdie told me Nov 15th for Rdna 3 We'll see.

 

[reb0rn]

I must quote our friend reb0rn here:



As can be seen, you sir have failed. Better luck next time, and I suggest that you use your brain .

pls be real how am I failed by powerpoint slides?

also only real word power use they compared to blended with 12900k and said only power per performance is duno was it 50-60% that mean nothing, cherry picked bench with old CPU

We all wait for 13900k vs 7950x temp and power use will be crucial not some stupid slides, but i already plan to get 7900x if I upgrade due AVX512 but gona wait real test not this crap today

 

[Vope45]

pls be real how am I failed by powerpoint slides?

also only real word power use they compared to blended with 12900k and said only power per performance is duno was it 50-60% that mean nothing, cherry picked bench with old CPU

We all wait for 13900k vs 7950x temp and power use will be crucial not some stupid slides, but i already plan to get 7900x if I upgrade due AVX512 but gona wait real test not this crap today

Just take the L and move on or the mods will get mad again.

 

[Det0x]

According to the press release, 6000 CL30. There weren't any scores, just relative performance uplift.

Could you please link the press release ?
I cant seem to find anything ;(

 

[nicalandia]

I think 15% over 5950x, over the 5800x3d might be too much.

5800X3D was on average 15% when compared to the 5800X. That was reduced about 9% when compared to the 5950X. The 7950X is about 15% gaming over 5950X.

The 7950X has just become the Gaming King....

 

[poke01]

even mobile

Zen 4 mobile will DESTROY Intel in efficiency and AMD laptops will have great battery life

 

[moinmoin]

The link from the YT video stream that contains the endnotes is still not working
https://www.amd.com/Ryzen7000disclosures

The endnotes not being part of the video is pretty stupid. Will I still get to all the fact checking before going to sleep?

 

[Thunder 57]

Best comment from the Youtube show:

"Someone better head over to UserBenchmarks house, like seriously they said they would do something if the 7600x matched the 12900k and...well yeah "

 

[inf64]

pls be real how am I failed by powerpoint slides?

also only real word power use they compared to blended with 12900k and said only power per performance is duno was it 50-60% that mean nothing, cherry picked bench with old CPU

We all wait for 13900k vs 7950x temp and power use will be crucial not some stupid slides, but i already plan to get 7900x if I upgrade due AVX512 but gona wait real test not this crap today

The slides are not "stupid", they carry legal repercussions if they are untrue. There is a chance the numbers are even a bit lower than what the reviews will show.

I hope you'll learn from this experience so you can be more humble in the future.

 

[CakeMonster]

Ian has a video up summarizing the news.

 

[tamz_msc]

Could you please link the press release ?
I cant seem to find anything ;(

Check ir.amd.com. I'm on mobile right now.

 

[Markfw]

The process isn't quite the same, and more cores still helps even at the same power limits. The slide Det0x just posted kinda demonstrates why.

7nm vs 5mn and a redesigned core ??? Lets talk later, but these slides demonstrate what I have been saying about power.

 

[Saylick]

29% Performance boost is mind boggling

Zen 5 will be interesting. They seem to be relying on cache and large frequency gains for these numbers. Which is great, but how much further can they push it for the next gen?

I think there's a 50-50 chance that Zen 4 ST gains might actually beat Zen 5's ST gains, as crazy as that may sound. I figure Zen 5 breaks the 6 GHz barrier with a >20% IPC uplift. To be honest, 20% is probably underestimating it, too, and is a rather "weak" guess. I think >25% IPC with a total ST perf increase of ~30% is within the cards given that Zen 5 is poised to be the BIGGEST architectural overhaul of Zen thus far.

 

[inf64]

The endnotes not being part of the video is pretty stupid. Will I still get to all the fact checking before going to sleep?

The Zen 4 page is up !

https://www.amd.com/en/processors/ryzen

Intel 12900K was tested with G.Skill DDR5-6000C30 (F5-6000J3038F16GX2-TZ5N)
Legal Disclosures for together we advance_PCs AMD Ryzen™ 7000 Video

1. Based on 5nm process node as of August 2022. 
2. Max boost for AMD Ryzen processors is the maximum frequency achievable by a single core on the processor running a bursty single-threaded workload. Max boost will vary based on several factors, including, but not limited to: thermal paste; system cooling; motherboard design and BIOS; the latest AMD chipset driver; and the latest OS updates. GD-150.
3. Video codec acceleration (including at least the HEVC (H.265), H.264, VP9, and AV1 codecs) is subject to and not operable without inclusion/installation of compatible media players. GD-176
4. Overclocking and/or undervolting AMD processors and memory, including without limitation, altering clock frequencies / multipliers or memory timing / voltage, to operate outside of AMD’s published specifications will void any applicable AMD product warranty, even when enabled via AMD hardware and/or software. This may also void warranties offered by the system manufacturer or retailer. Users assume all risks and liabilities that may arise out of overclocking and/or undervolting AMD processors, including, without limitation, failure of or damage to hardware, reduced system performance and/or data loss, corruption or vulnerability. GD-106.
5. Precision Boost Overdrive requires a compatible AMD Ryzen Threadripper, 3rd Gen AMD Ryzen processor, or Ryzen 5000 Series processor and a motherboard compatible with one or more of these processors. 2nd Gen AMD Ryzen processors including the Ryzen 3200G processor are not compatible with Precision Boost Overdrive. Because Precision Boost Overdrive enables operation of the processor outside of specifications and in excess of factory settings, use of the feature invalidates the AMD product warranty and may also void warranties offered by the system manufacturer or retailer. GD-135.
6. AMD’s product warranty does not cover damages caused by overclocking, even when overclocking is enabled via AMD hardware and/or software. GD-26.
7. Testing as of 15 August, 2022, by AMD Performance Labs using the following hardware: AMD Socket AM5 Reference Motherboard with AMD Ryzen™ 9 7950X, Ryzen™ 9 7900X, Ryzen™ 5 7600X and G.Skill DDR5-6000C30 (F5-6000J3038F16GX2-TZ5N) with AMD EXPO™; versus AMD Socket AM4 Reference Motherboard with Ryzen™ 9 5950X, Ryzen™ 9 5900X, Ryzen™ 5 5600X; versus ROG Maximus Z690 Hero with Core i9-12900K and G.Skill DDR5-6000C30 (F5-6000J3038F16GX2-TZ5N) with AMD EXPO™ loaded. ALL SYSTEMS configured with NXZT Kraken X63, open air test bench, Radeon™ RX 6950XT (driver 22.7.1 Optional), Windows® 11 22000.856, AMD Smart Access Memory/PCIe® Resizable Base Address Register (“ReBAR”) ON, Virtualization-Based Security (VBS) OFF. All games tested at 1920x1080 with HIGH in-game preset and the chronologically newest graphics industry API available within the game’s rendering engine (e.g. Vulkan® over OpenGL™, DirectX® 12 over DirectX® 11). Results may vary. RPL-007

 

[Zucker2k]

I think 15% over 5950x, over the 5800x3d might be too much.

I meant 5950x, which is the same figure for the 5800x3d over the 5950x. Which means Zen 4 = 5800x3d in gaming. Which means RPL should surely take the gaming crown until the Zen4 X3D is launched for the final showdown.

 

[Zucker2k]

That s 45% over the 5950X in Cinebench R23 since the 35% was related to this load, and 41% over the 12900K.

I didn't say CB. The whole argument was centered around efficiency. You wanted to eat your cake and still have it. You got everything wrong in your calculations.

Edit: Those multithreaded figures mean nothing if ADL is running at 125w and Zen 4 is running at 170w. We all know ADL reverts to base tdp on long runs or when it hits thermal limits.

 

[mikk]

AMD's been on a year and a half cadence, so that would make sense for them to target, if nothing else.

Almost 2 years between Zen 3 and Zen 4.

In the footnotes there are no power budget numbers from the CPUs, this is not good.

 

[poke01]

How big is core+L2 for the Apple M2?

Around the same as Zen4.