Wednesday, September 25th 2024
AMD Rushing in Ryzen 7 9800X3D, Expect Product Launch Late-October
Facing poor sales of its Ryzen 9000 "Granite Ridge" desktop processors, and with the spectre of Intel's Core Ultra "Arrow Lake-S" looming, AMD is rumored to have given its desktop processor roadmap a shakedown. The company is working to rush in at least one of the three upcoming Ryzen 9000X3D series processor SKUs. The Ryzen 7 9800X3D is a successor to the popular Ryzen 7 7800X3D. It pairs the new "Zen 5" microarchitecture with 3D V-cache technology to boost gaming performance. AMD is allegedly rushing the 9800X3D for a late-October launch. If this chip meets its performance targets (of around 15-20% over the 9700X), then AMD hopes it could take the edge off Intel's Core Ultra 200-series.
Launch of a Ryzen 9000X3D series product-stack became inevitable when AMD confirmed that the "Zen 5" CCD has silicon-level preparation for 3D V-cache (such as TSVs over the region with the on-die L3 cache that interface with the stacked L3D silicon), however, it was expected that the non-X3D Ryzen 9000 series, such as the 9700X, would perform close to the 7800X3D in games, giving AMD room to launch the 9800X3D in Q1-2025. Prior to the 7800X3D and Intel's 13th Gen Core "Raptor Lake," the Ryzen 7 7700X nearly matched the gaming performance of the Ryzen 7 5800X3D, and so something similar was expected of the 9700X. Of course things didn't go to plan, the 9700X fell significantly short of the 7800X3D in gaming, resulting in mixed reviews and low sales.The 9800X3D won't be the only chip from the 9000X3D series, there are also the Ryzen 9 9900X3D and new flagship 9950X3D planned, however, zhangzhonghao, the user behind this leak, says that the dual-CCD processors will do something different to the 7900X3D and 7950X3D to attract the class of buyers that wants both flagship gaming performance and productivity performance competitive to the Core Ultra 9 285K. The user did not elaborate on what these "new features" are, but if we were to guess, it's likely that both CCDs on the processor get 3D V-cache. The 9900X3D and 9950X3D are on-track for a Q1-2025 release.
Sources:
harukaze5719 (Twitter), zhangzhonghao (ChipHell forums), VideoCardz
Launch of a Ryzen 9000X3D series product-stack became inevitable when AMD confirmed that the "Zen 5" CCD has silicon-level preparation for 3D V-cache (such as TSVs over the region with the on-die L3 cache that interface with the stacked L3D silicon), however, it was expected that the non-X3D Ryzen 9000 series, such as the 9700X, would perform close to the 7800X3D in games, giving AMD room to launch the 9800X3D in Q1-2025. Prior to the 7800X3D and Intel's 13th Gen Core "Raptor Lake," the Ryzen 7 7700X nearly matched the gaming performance of the Ryzen 7 5800X3D, and so something similar was expected of the 9700X. Of course things didn't go to plan, the 9700X fell significantly short of the 7800X3D in gaming, resulting in mixed reviews and low sales.The 9800X3D won't be the only chip from the 9000X3D series, there are also the Ryzen 9 9900X3D and new flagship 9950X3D planned, however, zhangzhonghao, the user behind this leak, says that the dual-CCD processors will do something different to the 7900X3D and 7950X3D to attract the class of buyers that wants both flagship gaming performance and productivity performance competitive to the Core Ultra 9 285K. The user did not elaborate on what these "new features" are, but if we were to guess, it's likely that both CCDs on the processor get 3D V-cache. The 9900X3D and 9950X3D are on-track for a Q1-2025 release.
115 Comments on AMD Rushing in Ryzen 7 9800X3D, Expect Product Launch Late-October
While true, you acted as if you didn't know that we were talking about a 3D CPU. Such a theoretical CPU would need more L3, that's all I said.
The rest has been explained by others. You can't prove that. Working hard doesn't mean doing things right, they can still mess up in a thousand different ways. I'm not sure what you're talking about. Just how likely is it that more cache will give new issues? Obviosly not impossible, but I doubt it.
GCC (open source compiler), in my point of view, only utilises Ryzen 3000 or older properly. A bit can be gained by optimising the code
Let's hope it's not a one store deal.
I do not know where you looked or what you saw, but what I remember from what I saw was most of the blame being put on Microsoft for incompetence and with most people agreeing that Windows 11 in general as a whole is absolute garbage that is inferior to Windows 10. Again, it's the assumption that everyone is out for to put AMD down. At least we can agree that AMD Marketing blundered the release of Zen 5 by using over-exaggerated gaming benchmarks! Considering Intel's previous anti-competitive behaviour in the last 2 decades or so, I understand the animosity and suspicion, but do you have proof supporting this claim???
At least we can agree that the marketing team at AMD needs some heavy reform.
Ryzen 5 7600X3D tests reveal better performance than the 7900X3D - PC Guide
It's a guess tho
I don't use the pc just for gaming and the 7800x3d would have been a downgrade in MT, 7900x3d was an awful chip and the 7950x3d too expensive for what I needed. The 9700x at 175wppt matches my 5900x in MT in stock so a 9800x3d should be a decent upgrade in gaming and ST.
I hope AMD does something better with the 9900x3d but I am patiently waiting the 9800x3d to leave am4.
I'm not considering intel because my custom loop should work on AM5
.
At 350 and vs 400 I would in many cases chose 7950X3D, but where I live the difference is much bigger, 150usd atm. When launched the 7900X3D will be more than 50usd more expensive. Remember launchprices. Nowadays most want 7800X3D so it will cost more.
It was a lot worse on the original reviews. I would say, read between the lines.
The intel articles normally have positive headlines and you see a lot of them.
They keep giving them "the benefit of the doubt" or worse, paint them on a very positive light.
And with few exceptions, will ever mention any of intel illegal actions committed through the years or the many times they have lied on presentations, like hiding a chiller under the table to cool their CPU's.
AMD ones are usually the contrary, with a negative tone.
Given how things are today, you can't simply assume that all reviewers are really unbiased, everyone is in for the money. Indeed they do.
Why is it that you guys so often believe that most people have the secret goal of putting AMD down? Yes, the general perception of Zen 5 after the launch wasn't so positive, but the criticisms made were valid and were not done out of spite towards AMD. Yes, it was demonstrated that W11 was causing performance regressions for AMD CPUs, but even with the 24H2 fixes it still showed that in fact Zen 5 did not have that much of a major leap over Zen 4, and that in fact it is Microsoft that is to be blamed for W11 acting like trash because Hardware unboxed did a 42 game test with the 14700K with 24H2 comparing it to the 9700X and while not as much of an uplift like Zen 5, also showed performance uplift across the board, at least for gaming, so Intel CPUs were also being performance gimped due to W11 being trash. Yet for some reason there were so many people yelling "Wintel!"
This is what I mean, the underdog bias. It's admirable how AMD has managed to reverse and completely change everything with Zen, rising up and embarrassing Intel over and over, but that time has passed and now we have 2 competitive companies in the CPU market. Just because AMD was the underdog in the CPU market does not mean that when someone criticises them they are doing it out of malice or spite. There is no need to write something like "selective reading, selective rage, etc, only applies to AMD" to infer that most people were being critical only out of spite because it's AMD, because that's not true.
One example, everyone, even without trying it, will automatically call FSR trash, regardless of the version, regardless of some good samples or overall, its main "Pro" which is being platform agnostic. Thats part of the point, even now, everyone seems to automatically say that Zen 5 is absolute trash, which is not true. In my particular case, its not the underdog, but the fact that out of the big 3, they offer me the most pro consumer products and as they say, talk with your wallet, hence my preference. Indeed they have, but they are still not over the hill.
Intel has some big "allies", like Dell, which still refuses to use any of AMD chips on their lucrative business line, like the Latitudes laptops and Optiplex desktops. Sadly, many still do this out of malice. I wish that I was the only one thinking that the case, but I am not.
One of my favorites, when it was rumored that AMD blocked the use of DLSS in Starfield, Tim from Ngreedia Unboxed released 3 hit videos calling AMD all kinds of things.
When it was proven that they didnt do such request or action, he never apologized.
Or he has never criticized DLSS for being an anticonsumer tool which main existence is to keep you locked into that hardware.
Personally, I stopped watching them since that time.
Or when its the subject of power consumption, funny how it doesn't matter when its done by an Intel chip, but a travesty when its AMD.
But I think that we went far away off topic and we can conclude in agreeing with disagreeing.
Also ZEN 6 should increase the core count to 12 or 16 cores per CCD so if they release a 10800X3D chip with 16c/32t on 1 CCD with V-Cache that will be Day-1 for me!
I see no source claiming that it's rushed.
Make it symmetrically modular accessible on each half by CCX on each side. In the case of 4 CCX they'd probably want to basically take a corner frame access to the X3D cache slab if trying to work around the latency concerns. I think what they would likely do in that scenario is try to place the X3D at the center and then the 4 CCX dies around it in more of 5 sided dice arrangement and run traces diagonally to CCX die's accessing it. Like if they want the shorted trace lengths to the X3D at least. Like I can't think of a better way to go about it that doesn't involve stacked substrate and vertical height access between cache and chip die.