AMD previously told us that it's long-awaited Ryzen 9 3950X, a 16-core 32-thread behemoth destined for the mainstream desktop, would arrive in September 2019, but today the company announced it is delaying the release until November while it focuses on meeting the demand for existing chips.
The company did throw us a bone, though, and also announced for the first time that the third-generation Threadripper processors would launch in November, though the graphic clearly states they will debut with 24 cores instead of the expected 32, or even 64, cores.
Here is the official statement:
We are focusing on meeting the strong demand for our 3rd generation AMD Ryzen processors in the market and now plan to launch both the AMD Ryzen 9 3950X and initial members of the 3rd Gen AMD Ryzen Threadripper processor family in volume this November. We are confident that when enthusiasts get their hands on the world’s first 16-core mainstream desktop processor and our next-generation of high-end desktop processors, the wait will be well worth it.
Aside from the mention of 24 cores, AMD doesn't give us any specific details of the new Threadripper chips. There's no shortage of possible reasons the company has delayed the Ryzen 9 3950X, with the most obvious being the company's struggles to meet the current level of demand for its highest-end chips.
AMD's beastly 12-core 24-thread Ryzen 9 3900X is still a rarity at retail two and a half months after its launch, leading to price gouging from third-party retailers. While we're accustomed to scarce availability of chips from both Intel and AMD in the first few weeks of a launch, the persistent 3900X shortages are more pronounced than we've seen with other processors. That's led to plenty of speculation, with the most popular theories being that either AMD isn't yielding enough 7nm chips with high enough quality to satiate demand for the consumer chips, or that the company is setting the premium-quality compute die aside for its promising EPYC Rome data center processors, where demand is surely high due to the incredible blend of price, performance, and leading-edge features. In fact, the company just announced a newer, faster version of those chips two days ago, which would also obviously require cream-of-the-crop dies.
AMD's decision to use the same 7nm compute die for both versions of chips is a wonderful example of crafty reuse that allows it to realize impressive economies of scale while maximizing TSMC's production capacity, but it's logical to assume it could also serve as a constraint if there is booming demand for both types of chips. AMD has taken to a new binning strategy of offering chips with a mix of both fast and not-so-fast cores, which we discovered after the Ryzen 3000 series launch, which also maximizes production capacity. Even though lead times for new 7nm orders are rumored to be increasing at TSMC, it's rational to assume that AMD has accounted for any possible scenarios to deal with simultaneous demand, so the verdict is out if that is the only source of the delay.
AMD's pending launch may also be impacted by its focus on ironing out some of the issues with the Ryzen 3000 chips that are already on the market. AMD has come under withering fire from the enthusiast community due to the Ryzen 3000 series chips' inability to hit the rated boost speeds, which came to a boiling point with a recent survey conducted by YouTuber Der8auer in which a surprising number of respondents reported they have been unable to reach the rated boost frequencies with their Ryzen 3000 processors. Shortly thereafter, AMD announced that it would release a new firmware that improves clock rates by 25 to 50MHz, along with other fine-grained improvements. After several preliminary tests with beta firmware, we've found that AMD's updates correct the small clock rate deficit that we've seen with our Ryzen 3000 samples.
While the delayed 3950X launch is a disappointment for enthusiasts looking for a new level of performance on the mainstream desktop, AMD also announced that its Third-Gen Ryzen Threadripper processors would debut in November. This is encouraging given that AMD inexplicably removed those products from its 2019 roadmap earlier this year.
Today marks AMD's first official mention of a launch date for the new chips, though we've already seen plenty of leaked benchmarks that foretell their imminent arrival. It also appears that AMD will now split the Threadripper lineup into both workstation- and consumer-class tiers.
AMD's current-gen Threadripper processors come with the same underlying architecture that we see with the company's first-gen EPYC data center processors, and we expect the new third-gen models to follow that same trend. Given that the EPYC Rome lineup tops out at 64 cores, expectations are high that AMD will upset the HEDT space all over again with an amazing array of options. But it isn't a given that we'll see 64-core Threadripper processors, as that may be overkill for this relatively small market segment. However, we do expect AMD to at least match its previous-gen's 32 cores, but we won't know the final answer for a bit longer: AMD says the new chips will debut with 24 cores in November.
We'll update this post as more information becomes available.
The only problem I see is Intel is attacking HPC from many fronts and is offering many features that AMD might not have, like their FPGSa designed for Data Centers or Optane DIMMs. Its going to take a lot of work on AMDs end to really take a chunk out of Xeons and not just offering more cores per dollar.
I can't blame AMD for prioritizing Epyc if that's what they're doing. I know even my work, a long time Intel-only shop, has now started ordering Epyc servers due to Intel's nonstop parade of security issues. There will be growing pains as AMD expands their market share.
In addition to the memory channels (4 or 8), registered & ECC memory support (on the workstation parts) and PCIe lanes, there is another way they can differentiate between mainstream Ryzen and Threadripper: Number of CPU dies. A 3rd gen. Threadripper could potentially have anything between 1 and 8 of them.
By using partially disabled dies, but more of them, AMD could increase both the L3 cache amounts as well as memory bandwidth(*. Workstation users could use more cache and more bandwidth to feed the 8 channel memory.
24 core could be made from 3 dies of 8 cores, 4 x 6 or 6 x 4, which would yield max. 96/128/192MB of L3 cache, and 1.5x/2x/3x the memory bandwidth(* (compared to 64MB and 1x of a 2 CCD design such as 3900X or 3950X).
If instead of a 16-core (2x8 or 4x4), they decided to go for a 18-core, they could build it from 3 dies of 6 cores each. (As an added bonus, you'd be getting 2 "extra" cores compared to the mainstream CPU, and 18 is halfway between 12 and 24, if a 12-core Threadripper were to be the "starting at" model for people who just need I/O.)
*) between the CCXes and the I/O die
What I think would be interesting is if Epyc CPUs supported disabling half of their PCIe lanes and memory channels to slot into TR4 boards where people wanted > 32 cores. That would avoid AMD needing to create more TR SKUs, just to serve that small market segment.
I just think it'd be nice to see them take advantage of the abstraction provided by the I/O die, in order to unify their workstation & server chip markets. Create more options for everybody, without the burden of so many SKUs.