Monday, February 19th 2024
AMD "Zen 5c" CCDs Made On More Advanced 3 nm Node Than "Zen 5"
AMD is reportedly building its upcoming "Zen 5" and "Zen 5c" CPU Core Dies (CCDs) on two different foundry nodes, a report by Chinese publication UDN, claims. The Zen 5 CCD powering the upcoming Ryzen "Granite Ridge" desktop processors, "Fire Range" mobile processors, and EPYC "Turin" server processors, will be reportedly built on the 4 nm EUV foundry node, a slightly more advanced node than the current 5 nm EUV the company is building "Zen 4" CCDs on. The "Zen 5c" CCD, or the chiplet with purely "Zen 5c" cores in a high density configuration; on the other hand, will be built on an even more advanced 3 nm EUV foundry node, the report says. Both CCDs will go into mass production in Q2-2024, with product launches expected across the second half of the year.
The "Zen 5c" chiplet has a mammoth 32 cores spread across two CCXs of 16 cores, each. Each CCX has 16 cores sharing a 32 MB L3 cache. It is to cram these 32 cores, each with 1 MB of L2 cache; and a total of 64 MB of L3 cache, that AMD could be turning to the 3 nm foundry node. Another reason could be voltages. If "Zen 4c" is anything to go by, the "Zen 5c" core is a highly compacted variant of "Zen 5," which operates at a lower voltage band than its larger sibling, without any change in IPC or instruction sets. The decision to go with 3 nm could be a move aimed at increasing clock speeds at those lower voltages, in a bid to generationally improve performance using clock speeds, besides IPC and core count. The EPYC processor with "Zen 5c" chiplets will feature no more than six such large CCDs, for a maximum core count of 192. The regular "Zen 5" CCD has just 8 cores in a single CCX, with 32 MB of L3 cache shared among the cores; and TSV provision for 3D Vertical Cache, to increase the L3 cache in special variants.
Sources:
UDN, Wccftech
The "Zen 5c" chiplet has a mammoth 32 cores spread across two CCXs of 16 cores, each. Each CCX has 16 cores sharing a 32 MB L3 cache. It is to cram these 32 cores, each with 1 MB of L2 cache; and a total of 64 MB of L3 cache, that AMD could be turning to the 3 nm foundry node. Another reason could be voltages. If "Zen 4c" is anything to go by, the "Zen 5c" core is a highly compacted variant of "Zen 5," which operates at a lower voltage band than its larger sibling, without any change in IPC or instruction sets. The decision to go with 3 nm could be a move aimed at increasing clock speeds at those lower voltages, in a bid to generationally improve performance using clock speeds, besides IPC and core count. The EPYC processor with "Zen 5c" chiplets will feature no more than six such large CCDs, for a maximum core count of 192. The regular "Zen 5" CCD has just 8 cores in a single CCX, with 32 MB of L3 cache shared among the cores; and TSV provision for 3D Vertical Cache, to increase the L3 cache in special variants.
78 Comments on AMD "Zen 5c" CCDs Made On More Advanced 3 nm Node Than "Zen 5"
Again showing that testing large sample counts and averaging results gives more useful information than picking a single number and holding it up as supposed proof of something.
www.pcstats.com/articles/2582/3.html
Great argument.
So, with Zen 4 AMD increased wattage to 170W/230W.
You still keep throwing the ball outside the court.
I used to think that AMD would bring c core to the desktop as well, but mostly because Ryzen mid-range was suffering from the comparison against a 13600k/13700k. But the leaks so far, and them already reserving zen4c to the datacenter/low power laptop seems to suggest that as long as the Ryzen 9 are staying competitive, they are not interested to fight Intel on the core count. Even though their chiplet design makes this really easy. They would rather fight them with "real cores", and from what I've seen on various comment section or discord, AMD having more "real cores" is already a win marketing wise over intel's "cinebench accelerators". (Even though their mid-range using c-core would make them a better match against Intel, but I got called stupid for saying that, go figure! :D )
My general feeling is that the core count marketing advantage in 2024 is blown out of proportion, especially when the hottest, best-selling mainstream CPU is an 8 core that's more expensive than competing 12 or 16 cores CPUs. I've talked with "normies" back when Intel had a core count deficit, and they still liked Intel more because they just had more trust over seeing the brand more often on high-end SKU (hello dell XPS, alienware laptops etc... flagships are still being paired with an Intel CPU, even in the zen 3 era) From what I've seen normies are more sensitive to overall brand perception than shallow specs. That's how Apple manage to sell their computers, even though they always looked underpowered for the price.
I'm not arguing against you specifically, just the general perception that the e-cores are only there to accelerate benchmarks, padding the marketing material when they do bring tangible benefits irl. Yes it's a fix because Intel couldn't make smaller p cores, but that fix is doing what it's supposed to do : competing with AMD on performance rather than just being a pure marketing ploy to increase the core count on the spec list for cheap. But I'm also just a weird dude, who like exotic stuff, Ryzen chiplets already fascinated me just for being different from what Intel was doing for a more than a decade. Now I'm more interested to see where Intel is going with the exotic stuff that they've been pulling lately. If they hit a wall, that's that. I'm just going to buy AMD again.
Personally I can't wait for Zen 5 vs Arrow Lake tests.
AMD has its own special Windows scheduler to manage their issues with lopsided CCD performance in games. Special schedulers apply to both vendors.
AMD's driver is for user-level control of CPPC.
Funny.