• Welcome to TechPowerUp Forums, Guest! Please check out our forum guidelines for info related to our community.

AMD Ryzen 9000X3D Series to Keep the Same 64 MB 3D V-Cache Capacity, Offer Overclocking

Just my 2 cents, but does anyone even actually manually overclock Ryzens? :confused: PBO magic with undervolting/CO works great already.
At least one person out there does. I came across some reddit posts where a user supposedly sets the max voltage for the CPU based on the TSMC node (Zen3/Zen4) and then adjusts the frequency to whatever they can get away with. I've been too afraid to try that because 1) it came from reddit 2) I almost fried one CPU by accident already
 
We have reviews showing the 7950X3D at the exact same voltage and settings achieving higher efficiency when set to prefer cache as prefer frequency or stock:

View attachment 354489
View attachment 354492


Fetching data from cache as opposed to main system memory takes less energy and less time. Voltage is a factor in the efficiency but not the only factor.



Your claim was that it decreases performance:


Nothing on that claim eh?

As you put in bold, for games you want to be cache resident. The problem with the 7950X3D is that for certain games the OS places the game on the cores without the cache. Having cache on both would solve that issue, thus resulting in increased performance on the 7950X in select scenarios.



This is definitely not the whole picture as we know X3D has benefits outside of gaming (which was pointed out in the video)

Mind you we also know since that video was released that certain games do sometimes end up on frequency favored threads and those are the instances where performance would be improved. You can simulate the performance uplift the 7950X3D would see by using process lasso. It wouldn't exceed the 7800X3D's performance of course except for in games that use a lot of threads but it would bring the 7950X3D on par with the 7800X3D in games if not slightly ahead.



The 7000 series has 8 cores per CPU chiplet. This isn't relevant for the vast majority of games and applications.

You are also assuming that said cores need data from a different CCD. Having X3D on both CCDs is likely to increase the number of local cache hits. There is a reason AMD originally intended to use X3D for it's enterprise CPUs.



You you are assuming that a latency increase as a result of the the lower clocks is not more than offset by latency decreases from having a large fat cache stacked on the chip.

Your "basic logic" is drawing conclusions that aren't stated in any of the sources you provide as usual.
The main problem here is not only the CPU frequency, yeah the more the better.
At the moment if one thread needs to be transferred to another core on the other chiplet, all info about it needs to be transferred through Infinity Fabric. This operation needs time, so if you add here information from the x3D cache from CCD 1 to x3D CCD2 this will be huge as info but not a few bites.
Of course, if they can manage the x3D cache to be mirrored somehow this maybe will help, but I am not sure if this is possible.

So yeah there is a problem with 2 x3D chiplets when you need lower latency as in gaming.
 
Of course, if they can manage the x3D cache to be mirrored somehow this maybe will help, but I am not sure if this is possible.
LOL For some reason RAID X3D cache just popped in my head. Configure cache as RAID 0 or 1. Something new AMD could sell! Mirrored cache an interesting idea anyway. :rolleyes:
 
LOL For some reason RAID X3D cache just popped in my head. Configure cache as RAID 0 or 1. Something new AMD could sell! Mirrored cache an interesting idea anyway. :rolleyes:
Yeah but IF is used for all information from/to the system, so it must be very carefully managed, and de-facto is not infinite as the name lied :D
 
My 2022 Asus Stix I x670e motherboard has updated new ram kits up to 8 ghz tested.

fyi
 
I'm happy with single CCD CPUs for now.
Unless your use case changes you will likely be happy when 16-core chiplets come along with the "Zen 6 family".

LOL For some reason RAID X3D cache just popped in my head. Configure cache as RAID 0 or 1. Something new AMD could sell! Mirrored cache an interesting idea anyway. :rolleyes:
It would be amusing to see but ultimately pointless excepting some specific customised use cases and some ridiculous benchmark scores in some hyper specific test. Either way, it's worth pointing out that the L1, L2 and L3 caches on all AMD CPU's have been ECC since Zen 1, so somewhat analogous with RAID 1 as being there to avoid crisis.
 
Last edited:
Unless your use case changes you will likely be happy when 16-core chiplets come along with the "Zen 6 family".


It would be amusing to see but ultimately pointless excepting some specific customised use cases and some ridiculous benchmark scores in some hyper specific test. Either way, it's worth pointing out that the L1, L2 and L3 caches on all AMD CPU's have been ECC since Zen 1, so somewhat analogous with RAID 1 as being there to avoid crisis.
I believe the idea behind the cache mirroring was to avoid the cache miss if swapping between cores of different CCD's. Marketing would of course bastardize such a feature and call it RAID.
Redundant Array of Insanely-fast Data - I guess. (shrug) :shadedshu:
 
Just my 2 cents. After all of the noise I bought a 7800X3D. I was using a 7900X3D. It lasted for exactly 1 week. What I don't understand about the argument is that a 5900X is faster than a 5800X but a 7900X3D cannot be faster than a 7800X3D? I know that people are going to give me their opinions or post the 7600X3D numbers for a CPU that does not exist.
 
Just my 2 cents. After all of the noise I bought a 7800X3D. I was using a 7900X3D. It lasted for exactly 1 week. What I don't understand about the argument is that a 5900X is faster than a 5800X but a 7900X3D cannot be faster than a 7800X3D? I know that people are going to give me their opinions or post the 7600X3D numbers for a CPU that does not exist.
The difference is 5900x will use all the cores compared to 5800x. 7900X3D will only use the 6 X3D cores and park the other 6 cores where the 7800X3D will have 8 X3D cores available. But it's moot if your are 60fps gaming as all of the X3D CPU's are about on par with each other in that case if I recall correctly from looking at Hardware Unboxed videos.
 
Just my 2 cents. After all of the noise I bought a 7800X3D. I was using a 7900X3D. It lasted for exactly 1 week. What I don't understand about the argument is that a 5900X is faster than a 5800X but a 7900X3D cannot be faster than a 7800X3D? I know that people are going to give me their opinions or post the 7600X3D numbers for a CPU that does not exist.
It depends a little but yeah, with first bioses where 5800x boost with PBO up to 5.2GHz plus the benefit from one chiplet and it was faster in most games.
But AMD "fixed" this on both 5600x and 5800x to boost only 200MHz with PBO and then they become slower...
I gave my old 5600x that boost to 5.1GHz to my child and it still works perfectly, of course with old bios.
 
The difference is 5900x will use all the cores compared to 5800x. 7900X3D will only use the 6 X3D cores and park the other 6 cores where the 7800X3D will have 8 X3D cores available. But it's moot if your are 60fps gaming as all of the X3D CPU's are about on par with each other in that case if I recall correctly from looking at Hardware Unboxed videos.
In everything or just in Games that support V cache? Did HUB play TWWH3? That Game has no Vcache support. I game at 4K using a 144hz monitor.
 
In everything or just in Games that support V cache? Did HUB play TWWH3? That Game has no Vcache support.
To the best of my understanding from watching the HUB videos and reading others experience in TPU the AMD software/drivers (not sure which) with Windows Game mode is supposed to pin games to the X3D cache and park the other cores. I don't know how it decides which games it recognizes. You may have to manually intervene to get it to do something else especially for games that don't benefit from X3D.
I game at 4K using a 144hz monitor.
At a target of 60fps?
 
At a target of 60fps?
I cannot say for the game in question, but for example Skyrim is limited to 60fps at the game engine level and if pushed it simply makes the game crash.

Also, I set some (not all) games to run at 60fps minimum "target" via AMD "Radeon Chill", and then the limit is the monitor and the graphics card.
 
Back
Top