Thursday, January 5th 2023
AMD Confirms Ryzen 9 7950X3D and 7900X3D Feature 3DV Cache on Only One of the Two Chiplets
AMD today announced its new Ryzen 7000X3D high-end desktop processors to much fanfare, with availability slated for February 2023, you can read all about them in our older article. In our coverage, we noticed something odd about the cache sizes of the 12-core 7900X3D and 16-core 7950X3D. Whereas the 8-core, single-CCD 7800X3D comes with 104 MB of total cache (L2+L3), which works out to 1 MB L2 cache per core and 96 MB of L3 cache (32 MB on-die + 64 MB stacked 3DV cache); the dual-CCD 7900X3D and 7950X3D was shown with total caches of 140 MB and 144 MB, while they should have been 204 MB or 208 MB, respectively.
In our older article, we explored two possibilities—one that the 3DV cache is available on both CCDs but halved in size for whatever reason; and the second more outlandish possibility that only one of the two CCDs has stacked 3DV cache, while the other is a normal planar CCD with just the on-die 32 MB L3 cache. As it turns out, the latter theory is right! AMD put out high-resolution renders of the dual-CCD 7000X3D processors, where only one of the two CCDs is shown having the L3D (L3 cache die) stacked on top. Even real-world pictures of the older "Zen 3" 3DV cache CCDs from the 5800X3D or EPYC "Milan-X" processors show CCDs with 3DV caches having a distinct appearance with dividing lines between the L3D and the structural substrates over the regions of the CCD that have the CPU cores. In these renders, we see these lines drawn on only one of the two CCDs.
It shouldn't be hard for such an asymmetric cache setup to work in the real world from a software perspective, given that we are now firmly in the era of hybrid-core processors thanks to Intel and Arm. Even way before "Alder Lake," when AMD started shipping dual-CCD client processors with the Ryzen 3000 "Matisse" based on "Zen 2," the company closely collaborated with Microsoft to optimize OS scheduling such that high-performance and less-parallelized workloads such as games, are localized to just one of the two CCDs, to minimize DDR4 memory roundtrips.
Even before "Matisse," AMD and Microsoft confronted multi-threaded workload optimization challenges with dual-CCX architectures such as "Zen" and "Zen 2," where the OS scheduler would ideally want to localize gaming workload to a single CCX before saturating both CCXs on a single CCD, and then onward to the next CCD. This is achieved using methods such as CPPC2 preferred-core flagging, and which is why AMD highly recommends you to use their "Ryzen Balanced" Windows power-plan included with their Chipset drivers.
We predict that something similar is happening with the 12-core and 16-core 7000X3D processors—where gaming workloads can benefit from being localized to the 3DV cache-enabled CCD, and any spillover workloads (such as audio stack, network stack, background services, etc) are handled by the second CCD. In non-gaming workloads that scale across all 16 cores, the processor works like any other multi-core chip, it's just that the cores in the 3DV-enabled CCD have better performance from the larger victim cache. There shouldn't be any runtime errors arising from ISA mismatch, as the CPU core types on both CCDs are the same "Zen 4."
AMD Ryzen 7000X3D processors go on sale in February 2023.
In our older article, we explored two possibilities—one that the 3DV cache is available on both CCDs but halved in size for whatever reason; and the second more outlandish possibility that only one of the two CCDs has stacked 3DV cache, while the other is a normal planar CCD with just the on-die 32 MB L3 cache. As it turns out, the latter theory is right! AMD put out high-resolution renders of the dual-CCD 7000X3D processors, where only one of the two CCDs is shown having the L3D (L3 cache die) stacked on top. Even real-world pictures of the older "Zen 3" 3DV cache CCDs from the 5800X3D or EPYC "Milan-X" processors show CCDs with 3DV caches having a distinct appearance with dividing lines between the L3D and the structural substrates over the regions of the CCD that have the CPU cores. In these renders, we see these lines drawn on only one of the two CCDs.
Even before "Matisse," AMD and Microsoft confronted multi-threaded workload optimization challenges with dual-CCX architectures such as "Zen" and "Zen 2," where the OS scheduler would ideally want to localize gaming workload to a single CCX before saturating both CCXs on a single CCD, and then onward to the next CCD. This is achieved using methods such as CPPC2 preferred-core flagging, and which is why AMD highly recommends you to use their "Ryzen Balanced" Windows power-plan included with their Chipset drivers.
We predict that something similar is happening with the 12-core and 16-core 7000X3D processors—where gaming workloads can benefit from being localized to the 3DV cache-enabled CCD, and any spillover workloads (such as audio stack, network stack, background services, etc) are handled by the second CCD. In non-gaming workloads that scale across all 16 cores, the processor works like any other multi-core chip, it's just that the cores in the 3DV-enabled CCD have better performance from the larger victim cache. There shouldn't be any runtime errors arising from ISA mismatch, as the CPU core types on both CCDs are the same "Zen 4."
AMD Ryzen 7000X3D processors go on sale in February 2023.
164 Comments on AMD Confirms Ryzen 9 7950X3D and 7900X3D Feature 3DV Cache on Only One of the Two Chiplets
Edit: Here, my CPU's preferred cores are #8 and #2.
So:
7700X3D 120W (8 cores, lower boost frequencies)
7950X3D 120W (16 cores, higher boost frequencies)
In the real world, they will consume more than the 7700X and 7950X, respectively
The entire 7000X3D range is marked with 120W Default TDP, but we will see the real consumption in the reviews.
As for 7000X versus 7000X3D, the consumption of that extracache is added to the total consumption.
So, the X3D variants:
1. They will consume more than the X variants.
2. They will consume the same or less, in which case the frequencies will be lower than in the X versions, with the performance penalty that the 5800X3D also suffered in front of the 5800X, with the exception of games.
Update the base clocks are also significantly lower at 11% to 6% respectively.
Regarding the 7950X/7950X3D consumption, the answer was readily available: the X3D is clocked at 4.2 GHz base, 300MHz less. This is the only way to maintain the same consumption as the X counterpart. All will follow the trajectory of 5800X3D: increased performance in gaming with penalty in other software.
if task B depends on the result of task A, with task B running on a ccd with 3d cache, and task A runs on the ccd without 3d cache, then you'll get worse performance than if it was the other way around (provided task B takes more time to do than task A)
if there are enough threads and the scheduler isn't dumb, then sure you wouldn't notice a thing, but I would be surprised if, for example, a 7600x3d performed better than a 7900x3d (or if locking game threads to the ccd which does have the 3d cache) since there would be no potential for improper thread assignment
I think 7900x3d and 7950x3d will be a case of "more fps than 7900x, with extra performance to be had if you limit games to use the 3d cache ccd only"
probably a miniscule amount of latency, but I really wouldn't be surprised if locking games to use the 3d cache ccd would yield better performance than using both
at the very least it sounds like a better idea than what intel is doing with their p+e cores anyway
At this point, it'd just be low-hanging fruit, and could make use of bad bins of X3D-capped CCDs intended for the 5800X3D for "5600X3D" and "5700X3D", and conversely, if bins are already very good, use the best for 5950X3D and second best for 5900X3D. It wouldn't be much different than AMD having also used some Zen+ or Zen2 CCDs for refreshed 1#00 and 2#00 Ryzens that they were originally still producing in small batches for the budget market (which was a minor but welcome surprise when it was known), and it'd be a nice bone to toss towards the AM4 crowd.
7950X3D: 4.2GHz base clock
It will lose like the 5800X3D lost. It does not matter how much. The price matters and if it will bring you something extra in gaming. I doubt it will be worth it if you don't have a top video card.
It is vital for AMD to regain the crown in gaming. We are talking about some frames, impossible to detect in real life but with an impact in business. For the customer, hmmm, advertising, the soul of commerce.
Which part is unclear to you? I can help you. Unlike you, who made a fix for me, I will also use arguments.
Update the Tjmax was lowered by 6 degrees to 89* C
videocardz.com/newz/amd-ryzen-7000x3d-cpus-can-be-distinguished-by-a-new-box-lower-tjmax-temperature-confirmed
Something like: 5800X versus 5800X3D with:
3050, 3060 ... 3070 and/or 6500XT, 6600 ... 6700XT
How much does the X3D version add to these weaker video cards and IF IT'S WORTH IT (higher price, lower performance in many applications).
Something like: 4090 versus 4080 with:
intel pentium 4
How much does the 4090 version add to these weaker processors and IF IT'S WORTH IT (higher price, lower performance in many applications).
Latest TPU review. Enjoy your reading!
However, it is impossible to find (a random example) how much X3D helps an RX 6600 or RTX3060 compared to the X counterpart. :confused: