• We’re currently investigating an issue related to the forum theme and styling that is impacting page layout and visual formatting. The problem has been identified, and we are actively working on a resolution. There is no impact to user data or functionality, this is strictly a front-end display issue. We’ll post an update once the fix has been deployed. Thanks for your patience while we get this sorted.

Discussion Zen 5 Speculation (EPYC Turin and Strix Point/Granite Ridge - Ryzen 9000)

Page 862 - Seeking answers? Join the AnandTech community: where nearly half-a-million members share solutions and discuss the latest tech.
So if 9800X3D reviews drop in 1 week, when will they announce that thing? I mean there's still no official specs, AMD just posted 7th of November, so the CPU will just drop that day without an announcement before?
 
Why are you want to use cudimm on zen5? On zen5 cudimm only going to work at bypass mode, the ckd is not working. It’s wasting money, the first cudimm will cost like 400usd.
Agreed. RAM tuning has its merits, but it's a hard sell for me if I could have upgraded to a whole new tier of CPU for the same added price as the more expensive, more exotic RAM.
 
Agreed. RAM tuning has its merits, but it's a hard sell for me if I could have upgraded to a whole new tier of CPU for the same added price as the more expensive, more exotic RAM.
RAM tuning is so annoying to me that I just don't tune. Too many little buttons and cifers.
A one button push does it all would be cool lol.
 
RAM tuning is so annoying to me that I just don't tune. Too many little buttons and cifers.
A one button push does it all would be cool lol.
Luckily someone like buildzoid make guides where you can simply put in his numbers and make them work with your system if you have the same memory-kit (die) and platform, which makes it super easy. But besides that, I don't tune my memory either.
 
amd need to change the io die to support cudimm, I think need to wait till zen6.
And I think the final form of cudimm should be camm2, zen6 should be using it.

My original thought was that it would require a new IO die as well, however some have said that both Zen 4 and Zen 5 have the capability, but lack the software, par for the course with AMD, if true, so I felt I would bring both items up here. 🤣
 
My original thought was that it would require a new IO die as well, however some have said that both Zen 4 and Zen 5 have the capability, but lack the software, par for the course with AMD, if true, so I felt I would bring both items up here. 🤣
zen4 and zen5 can use cudimm, but it’s only going to work at bypass mode. Basically it will become a udimm.🤣totally wasting money.

Now only ARL can fully support cudimm, but ARL’s gaming performance is suck. So, no point to spend money on it too.🤣
 
zen4 and zen5 can use cudimm, but it’s only going to work at bypass mode. Basically it will become a udimm.🤣totally wasting money.

Now only ARL can fully support cudimm, but ARL’s gaming performance is suck. So, no point to spend money on it too.🤣

With the volume of ARL shipping (likely miniscule), it is hard to see how this memory standard can become a success. High price and almost no demand....

Which also brings up the mobo makers, and their ARL mobos. They are going to take a bath, losing a lot of money they sunk into these ARL motherboards.

But it will be fully deserved. Karma is finally gong to catch up with them.
 
With the volume of ARL shipping (likely miniscule), it is hard to see how this memory standard can become a success. High price and almost no demand....
So what you are basically saying is CUDIMM is RDRAM redux, with only supporting platform performing badly and needing CUDIMM to perform anywhere near decently, while the competition that performs better doesn't need one.

Got it.
 
If you think a 30$ price hike is too much for unrivaled #1 gaming performance, i guess you can wait 1 or 2 years for it to drop below 400$, same as with the 7800X3D, which btw have been selling like hot cakes from its very release.

I find the price pretty reasonable, especially when my next GFX card is expected to cost over 2 grands (5090)
 
Last edited:
Last edited:
Why? Isn't one of the 3D advantages that you get higher performance even with not that fast RAM?
Exactly. It's pointless. Could help on standard zen if other bottlenecks were removed but it's also pointless there with those bottlenecks. It's a similar argument to wanting more cores on the mainstream platform while ignoring the bandwidth bottlenecks. All pointless (given other constraints)
 
Except it doesn't because of other bandwidth limitations and infinity fabric. It doesn't really help until those things are addressed
Let's suppose it doesn't help in bandwidth. There's still the possibility of running the CUDIMM at lower latencies at 6400 MT/s, like CL26 or even lower. The stabilized signal integrity of CUDIMM should help there.
 
Back
Top