Latest AMD Chipset Driver 5.01.29.2026 WHQL Adds 3DV Cache Optimization 1.0.0.7

[btarunr]

AMD is preparing to release the latest version of its Chipset software. Version 5.01.29.2026 WHQL should be of particular interest for users planning to buy an upcoming Ryzen 7000X3D series processor. The chipset driver includes the "3D V-Cache Performance Optimizer driver" component, version 1.0.0.7. This gives Windows a degree of awareness of the asymmetric nature of 3DV cache on the 16-core 7950X3D and 12-core 7900X3D, where only one of the two CCDs (chiplets) has the 3DV cache memory, while the other is a regular "Zen 4" CCD with 32 MB on-die L3 cache. This awareness should in theory improve performance in less-parallelized workloads (such as games). The AMD website doesn't yet list this driver, but it should appear as we head closer to the market-availability date of the 7950X3D and 7900X3D (February 28). From the looks of it, the optimization works for both Windows 11 and Windows 10, so those on the older operating system for reasons, can keep rocking it.



View at TechPowerUp Main Site | Source

 

[VulkanBros]

And why not for the 5800X3D? Is it a ZEN4 thing....

 

[kapone32]

And why not for the 5800X3D?

if it's not broken don't fix it baby.

 

[PrEzi]

And why not for the 5800X3D? It is a ZEN4 thing....

Well duh! Because it only has one CCD which is already populated with 3D Cache? This driver also won't be necessary for the 7800X3D because it will follow the same principle (one VCache CCD only).
It is ONLY important for 2 and more mixed CCD architectures, where not all of them have VCache on them -- like 7900 or 7950 with 3D Cache (where only one CCD has 3D Cache and one doesn't).

 

[Denver]

And why not for the 5800X3D? Is it a ZEN4 thing....

"The chipset driver includes the "3D V-Cache Performance Optimizer driver" component, version 1.0.0.7. This gives Windows a degree of awareness of the asymmetric nature of 3DV cache on the 16-core 7950X3D and 12-core 7900X3D, where only one of the two CCDs (chiplets) has the 3DV cache memory, while the other is a regular "Zen 4" CCD with 32 MB on-die L3 cache."

Looks like you just read the title.

 

[robert3892]

ASRock already has this driver available for the X670E Steel Legend. Their BIOs is also updated.

 

[mechtech]

Interesting

 

[phanbuey]

Im really excited to read the reviews of these. It can really go either way. We all know the 7800x3d will be good, but will be exciting to see AMD use disparate cores for the first time. Could be great, could be a $#(*&Y show, will be super interesting in either case.

 

[Lucas_]

And why not for the 5800X3D? Is it a ZEN4 thing....

its 2 ccu against 1 , for the 5800 no need to priotize anything on the one which has more cashe.

 

[HD64G]

Only thing needed was that driver to allow the OS to communicate with the updated UEFI and in turn with the CPU in order to prioritize the use of the CCD that doesn't have the 3Dcache on it.

 

[Godrilla]

Only thing needed was that driver to allow the OS to communicate with the updated UEFI and in turn with the CPU in order to prioritize the use of the CCD that doesn't have the 3Dcache on it.

It will be interesting to see how much latency is added when using the faster clocked ccd and the extra cache from the second ccd if this even an option vs just using the the one extra cache ccd in gaming.

 

[zlobby]

Why on Earth that's not a part of UEFI ucode or Windows-delivered drivers?

 

[Zubasa]

Why on Earth that's not a part of UEFI ucode

There is.

https://twitter.com/i/web/status/1627974532282875904

 

[HisDivineOrder]

Remember when I read people saying how they just wanted an uncomplicated CPU that did what it did without fancy tricks? They said Alder Lake and Raptor Lake were just too complicated and using scheduler tricks was just downright uncivilized. Well, I guess they're really upset now that everyone's doing the overcomplicating.

 

[A Computer Guy]

I wonder how the 3D v-cache effects virtual machines? Is there any advantages there or just predominantly in Gaming?

 

[zlobby]

There is.

https://twitter.com/i/web/status/1627974532282875904

Yeah, but I expected AMD to streamline this a bit. After all, how many people care about chipset driver changelogs?

 

[AnotherReader]

I wonder how the 3D v-cache effects virtual machines? Is there any advantages there or just predominantly in Gaming?View attachment 285188

It would depend upon what the VMs are being used for, but in general, I wouldn't expect any advantage. It isn't just gaming though; any application with a working set larger than the 32 MB of the regular CCD would benefit.

 

[tabascosauz]

Yeah, but I expected AMD to streamline this a bit. After all, how many people care about chipset driver changelogs?

Probably still need some software element to make Windows scheduler function properly. Remember the 7950X's Win 11 gaming regressions due to loads not being confined to CCD1 properly? Same deal except the stakes are even higher now.

The "OS agnostic" remark is pretty amusing to me - for 3 years Ryzen at every turn has been heavily dependent on complementary Windows optimizations just to work properly, and now according to random Twitter dude it's suddenly "OS agnostic"?

As for the Core Flex feature, I'm not fully convinced yet that it's an AMD-wide feature contained in AGESA. We've seen proprietary features under the Tweaker page before (remember the dynamic OC switcher on Dark Hero that would "automatically" toggle between all-core and PBO depending on load?). I don't believe for a second that AMD will solve this problem through AGESA alone, without any help from Microsoft or chipset drivers.

 

[Punkenjoy]

Remember when I read people saying how they just wanted an uncomplicated CPU that did what it did without fancy tricks? They said Alder Lake and Raptor Lake were just too complicated and using scheduler tricks was just downright uncivilized. Well, I guess they're really upset now that everyone's doing the overcomplicating.

This is still way simplier than intel Big.Little solution. Both CPU have the same instruction set and architecture and in this case is more optimisation of performance than optimisation of die size.

The goal with dual CCD is to have 1 that clock very high if you run somthing that aren't cache sensitive, and one that have a lot of cache if you run something that does. If you happen to run something on the wrong CCD, you still run it on a fast Zen 4 core anwyay.

Intel on the other side, would have a really hard time if games would run on E-Cores by mistakes.

But in the ends, as long as the scheduler is doing it's job properly, this is not a big deals. The main drawback in my opinion of Intel E-cores is the fact they had to disable AVX-512 to maintain the same instruction set across all cores.

 

[A Computer Guy]

Probably still need some software element to make Windows scheduler function properly. Remember the 7950X's Win 11 gaming regressions due to loads not being confined to CCD1 properly? Same deal except the stakes are even higher now.

The "OS agnostic" remark is pretty amusing to me - for 3 years Ryzen at every turn has been heavily dependent on complementary Windows optimizations just to work properly, and now according to random Twitter dude it's suddenly "OS agnostic"?

Yea that is kind of funny when you put it that way.

As for the Core Flex feature, I'm not fully convinced yet that it's an AMD-wide feature contained in AGESA. We've seen proprietary features under the Tweaker page before (remember the dynamic OC switcher on Dark Hero that would "automatically" toggle between all-core and PBO depending on load?). I don't believe for a second that AMD will solve this problem through AGESA alone, without any help from Microsoft or chipset drivers.

How does Linux handle these core issues?

 

[HD64G]

It will be interesting to see how much latency is added when using the faster clocked ccd and the extra cache from the second ccd if this even an option vs just using the the one extra cache ccd in gaming.

Since the updated UEFI will have features to allow anyone to circumvent the default settings about the CCD selection depending on the app, we will see many tests about that and get a clear view in a few days.

Remember when I read people saying how they just wanted an uncomplicated CPU that did what it did without fancy tricks? They said Alder Lake and Raptor Lake were just too complicated and using scheduler tricks was just downright uncivilized. Well, I guess they're really upset now that everyone's doing the overcomplicating.

It will be simple to get >95% of the performance without messing with settings. But as always, the enthusiasts will get more using some new features and algorithms with some more time invested in it. On the Intel side, either you keep the E-cores on or you disable them depending on the apps you use. Nothing is very complicated if you do not want to seek the last few % of performance.

 

[A Computer Guy]

Since the updated UEFI will have features to allow anyone to circumvent the default settings about the CCD selection depending on the app, we will see many tests about that and get a clear view in a few days.


It will be simple to get >95% of the performance without messing with settings. But as always, the enthusiasts will get more using some new features and algorithms with some more time invested in it. On the Intel side, either you keep the E-cores on or you disable them depending on the apps you use. Nothing is very complicated if you do not want to seek the last few % of performance.

It would be great if there was a provided means to program threads to specifically run on P vs E cores or cached CCD vs. non-cached CCD's based on their capabilities.

 

[davidhhh]

Well, it's Feb 28, and the x3d chips are out, and still no drivers on the AMD site.....

 

[HD64G]

It would be great if there was a provided means to program threads to specifically run on P vs E cores or cached CCD vs. non-cached CCD's based on their capabilities.

The new driver with the updated UEFI allows that selection as @W1zzard explained thoroughly in his review on the 7950X3D.

 

[Klemc]

IDK :

2023/28/2 AMD

https://www.amd.com/en/support/chipsets/amd-socket-am5/b650e

is equal to ? :

2023/3/01 Asus

ROG STRIX B650E-E GAMING WIFI | Gaming motherboards｜ROG - Republic of Gamers｜ROG France

The ROG Strix B650E-E Gaming WiFi is a future-ready B-series beast, featuring 16 + 2 power stages, hyperspeed DDR5 memory, and next-gen PCIe® 5.0 connectivity. It features four M.2 slots with heatsinks, USB 3.2 Gen 2x2 support, WiFi 6E, AI Cooling II, AI Networking, Two-Way AI Noise Cancelation...

rog.asus.com