Overclock.net banner

[Official] Zen 5 X3D Owner's Club (9800x3D / 9900x3D / 9950x3D)

8 reading
1M views 10K replies 534 participants last post by  Luggage  
#1 ·
Image


Just like the Zen 4 X3D thread, I’m opening up this discussion prior to the anticipated launch/announcement of this generation’s 3D vcache offerings. Until launch and official announcement from AMD, everything posted here should be taken with a heavy grain of salt. Until then, rumors, predictions, theory-crafting, and general discussion is encouraged. I'll start the thread out with a collection of early details, and will update with more up and through launch, after-which this thread will transiton to a genral discussion between current and prosepective Zen 5 X3D owners, focusing on performance optimization, troubleshooting, and experience exchange.

9800x3D (8 cores):
Rumored Announcement Date:
October 25, 2024
Rumored Launch Date: November 7, 2024
Expected MSRP: $499 USD
Base clock: 4.7ghz
All core boost clock: 5.2ghz
Single core boost clock: 5.5ghz

9700x3D (12 cores) / 9950x3D (16 cores) annnouncment, launch dates, specs and pricing currently unknown.

Early Benchmarks:
Image


Image


Image


Image


Cinebench 2024 5.2ghz all-core boost:
Image
 
#2 ·
Have a feeling the boost clk for 9800X3D will be no more than 5.2GHz before OC.
 
#3 ·
Hard to agree. CBR24 is all AVX2/AVX512, and the screenshot there shows 5.2GHz. As of now, a 7800X3D would peak at 4.8GHz in CBR24. So I may speculate that the actual peak freq in normal loads, non AVX2/AVX512, will be more like 5.4-5.5GHz
 
#11 ·
Several updates this morning:

AMD China hosting a 2 day event from Oct 23-24 centered on ‘AMD’s new support for gaming experience.’ The invites feature the new Zen 5 x3D packaging, and the event is clearly timed to overshadow the Arrow Lake launch.

Image


Image


Official announcement will either coincide with event or occur the day after, on Oct 25.

Memory OC on the x3D child, paired with x870 motherboards looks extremely promising:
Image


Gigabyte announces Zen 5 x3D specific BIOS OC feature:


Image
 
#13 ·
#14 ·
#28 ·
#18 ·
Those AI or one button overclocking never work as well as manually tuning yourself since it has to take in a lot of variables. Though I could be wrong here who knows.
 
  • Rep+
Reactions: RedF and J7SC
#19 ·
...yeah, Asus has the NitroPath and Gigabyte that new Turbo thing (which will work on 7950X3D/ 670E as well :))...logically, MSI and ASRock should follow suit with their own press releases - trying to excite the crowd after a so-so R9 market intro and sales. Best to wait and see how these new tech will perform in real-world conditions. I look forward to it, though.
 
#20 ·
First, ASRock has to fix that:
You can't go below VDDQ 1,44V if you are in HVM (above 1,43V) with VDD
Image


Answer from ASRock Support:

"In the OC scenario, the VDD and VDDQ voltages must be set to the same value. Setting them to different values would lead to unstable or unexpected behavior. It is advisable to set both VDD and VDDQ to the same value. We are not sure why the products of other brands do not show the same behavior, but from a technical support point of view this is not a problem.":poop:
 
#21 ·
So what's up with this result from HUB? I can see the cheap boards dropping clocks, but not this much variation in the higher end boards. Differences aren't huge, but still I would think they would be the same?

Image


This is interesting too. He only ran 1 hour of P95, but gives a decent indication of which boards may currently be capable.

Image


 
#25 ·
#32 · (Edited)
This 9800X3D chip should be a beast, can’t wait to see what it does to the 285K. I haven’t ran an AMD system since the olden days, that just might change. I’d probably be more inclined towards the 9950X3D personally, but either one is going to be great.
 
  • Rep+
Reactions: bmagnien
#33 ·
#34 · (Edited)
#36 · (Edited)
My guess is that they reused the same mechanism that, on X3D models, forces a videogame to run on the single X3D CCD (that is, an automatic "game mode" power profile signal that gets sent to the firmware that changes the CPPC availability for that specific process only, when in foreground) to disable SMT and force the task execution to a single CCD for non-X3D models too.

That would reliably provide "some" performance improvement to a few videogames that don't need more than 6-8 cores at the same time. But a 35% delta may hardly be reached in that way (the average game fps may improve at most by a 5% only with SMT off on a 7800X3D) so they may have also added some further power-states optimisations thanks to the lower power requirements when disabling SMT.

I tend to agree with Melan that if they went down the generic automatic ram/fclk/cpu freq oc path instead, it would just cause instability to random cpu bins and be counterproductive to their brand, so I don't think they did that.

Actually, if you have a 9950X, you can validate my guess on this improvement by running some videogames benchmarks while using ProcessLasso to force execution on a single CCD only, and also disable the SMT-adjacent core on that CCD.
 
#35 ·
#37 · (Edited)
This is a 30% improvement over the 19.5k that I’d consider edge of stability CO oc for the 7800x3d in cb23 mt (without BCLK oc). That is…awesome?

+ the hwinfo SS from that same run just to keep everything in one place
Image
 
#42 ·
Also, I guess AMD was already doing this, but it feels like they might be further segmenting the non-x3D and x3D CPUs.

I wonder if in the end Zen5 will make sense with good non-gaming gains for normal Zen5 and good gaming gains for x3D. If true, makes a good range of offerings.
 
#43 ·
The price increase I can some what agree with.(over 500 dollars for 9800X3D) AMD know Intel have no match for gaming performance.
 
#45 · (Edited)
The reason why the 4080 from ju-rek and the 4090 from bmagnien provided roughly the same score is because at that low resolution the GPU is not the bottleneck. That means that it is a good test to evaluate the relative CPU performance difference with the Zen 5.
 
#46 ·
#62 ·
My daily 7800x3D: FINAL FANTASY XIV: Dawntrail Official Benchmark

Windows is latency optimized, minimal background, display res @ 1280x720.

Realistic stable uplift is likely 15%, not full 21%. Leak is using bclk and likely cold or chilled water and is unlikely to be 24/7 stable.
 
  • Love
Reactions: Antsu
#65 · (Edited)
Dawntrail appears to be a Zen5 outlier. One of the few games that saw big gains.

Still impressive result, but I wouldn't expect this across the board.

7950x: 291.6 fps
9950x: 336.2 fps

15% gain

Toss in another 2-3% for the rumored clock speed increase and you've got 17-18% which is approximately what that leaked 9800x3D Dawntrail run suggests given the highest 7000x3d score I've seen is 53k (Anandtech forum).

Edit: The 53k result on Anandtech might be a little inflated relative to other scores as that user is using a 6900XT. AMD drivers are known for having less overhead.

Image


 
#68 ·
Image