The arrival of AMD’s 1000-series Ryzen processors (like the flagship Ryzen 7 1800X) revitalized the CPU market in ways we couldn’t have imagined just a few years ago, forcing Intel to become more competitive across the full breadth of its product stack. But that was just the beginning of AMD’s assault with its new Zen-based chips. Last year AMD moved from its original 14nm chips to the 2000-series 12nm Ryzen processors (like the Ryzen 7 2700X) that come wielding the Zen+ microarchitecture.
These newer chips were a solid evolutionary step, but now AMD is working feverishly on bringing its Zen 2 microarchitecture to market packing a 7nm manufacturing process. Along with an improved Infinity Fabric, the collection of new technologies will come to market as the third-generation Ryzen ‘Matisse’ processors. If AMD can pull off this feat in a timely manner while Intel still struggles to roll out its 10nm CPUs, it will mark the first time in its history that AMD has taken the process node lead from Intel.
AMD recently demoed the new processor during its CES keynote. While we still don’t know all the details about what is coming with AMD’s 3000-series processors yet, we do know they are coming to market in mid-2019. Let’s take a look at what we do know, and what we don’t.
Third-Gen Ryzen Performance
At CES 2019, AMD CEO Lisa Su showed off an “early” third-gen Ryzen desktop processor running a game paired with a Radeon VII. These 7nm chips will be the first desktop processors to support PCIe 4.0 x16 right out of the box. That also implies that the new 500-series chipsets will also support the new standard.
AMD demoed a pre-production Ryzen 3000-series processor going head-to-head against a Core i9-9900K. It’s noteworthy that AMD hasn’t fully tweaked the design yet, meaning that it will likely extract more performance from the chip before it comes to market. The eight-core 16-thread Ryzen processor was about even with the eight-core 16-thread Core i9-9900K in a Cinebench multi-threaded workload, with the Ryzen score 2,057 to the i9’s 2040.
Third-gen Ryzen’s power consumption figures during the demo were just as impressive – the chip consumed 30% less power than the Intel processor during the benchmark. Intel’s Core i9-9900K requires a high-end motherboard, power supply, and cooler to extract the optimum level of performance. The Ryzen processors’ comparatively low power consumption means it could be much cheaper to build full systems around the chips, thus allowing AMD to offer a similar level of performance while maintaining the value advantage.
We analyzed AMD’s test notes from the show, and both test systems seem evenly matched. AMD also topped both processors with the capable Noctua NH-D15S cooler, so it’s unlikely that Intel’s Core i9-9900K was hampered by poor thermals.
The 3rd-Gen Ryzen Chip
As expected, the third-generation Ryzen processor comes with a multi-chiplet arrangement, much like the EPYC Rome processors AMD recently announced. This modular design consists of an eight-core 7nm chiplet (fabbed at TSMC) connected to a 14nm I/O die (fabbed at GlobalFoundries). The central chip ties connects the eight-core die to the PCIe, memory, and I/O resources. Aside from support for PCIe 4.0, AMD is still keeping details about these resources close to the chest.
As we can see, AMD offset the larger I/O die from the 7nm core die, leaving plenty of room for a second eight-core CPU complex. We can even see what appears to be a dip in the PCB (images below) that would serve as a mounting point for another eight-core chiplet. AMD’s Lisa Su has commented that the design obviously leaves room for another chiplet, and that “you might expect that we will have more than eight cores.” However, Su did not specify how many more cores will come with the processors, or if they would arrive during the initial launch or come as an iterative follow-up series of chips.
AMD is now using a second-gen Infinity Fabric to connect the multi-chip design with a 14nm I/O die that serves as the linchpin of the design. This design helps AMD keep the areas of the chip that don’t scale well, like the memory controllers and I/O, on a proven and mature node, while also leveraging the performance, density and economic advantages of the 7nm node for the important compute functions. The innovative design is a sign of a broader trend in the industry to heterogeneous architectures.
PCIe 4.0 Support
AMD led the way with PCIe 4.0 support for its 7nm EPYC Rome chips and enabled the standard on the 3000-series processors. However, AMD has confirmed to us that while the new Ryzen processors are compatible with motherboards designed for the first-gen chips, it will require new motherboards to fully support the PCIe 4.0 standard.
There is hope for existing Ryzen systems, though. AMD representatives confirmed that 300- and 400-series AM4 motherboards can support PCIe 4.0. AMD will not lock the out feature, instead it will be up to motherboard vendors to validate and qualify the faster standard on its motherboards on a case-by-case basis. Motherboard vendors that do support the feature will enable it through BIOS updates, but those updates will come at the discretion of the vendor.
Most older motherboards could support a PCIe 4.0 x16 connection to the first slot on the motherboard, but the remainder of the slots could revert to PCIe 3.0 signaling rates. That’s because any trace routing on the motherboard that exceeds six inches requires newer redrivers and retimers that support PCIe 4.0’s faster signaling rates. That means the PCIe slot nearest to the CPU will easily support PCIe 4.0, while the other slots, including M.2 ports, could run at a PCIe 3.0 signaling rate. Support could be limited to slots based upon board, switch, and mux layouts.
The 500-Series Chipset
The third-gen Ryzen processors support PCIe 4.0, and it’s logical to assume that the next-gen chipsets, largely thought to be named the 500-Series, will also support the standard. Our sources tell us the 14nm 500-Series chipsets will consume more power (~15W) than the 28nm chipsets (~8W) used on current AM4 motherboards, but that’s because the 500-series chipsets also support PCIe 4.0. We weren’t told the specific lane allocations of the new chipset, but those faster lanes will be useful for numerous types of secondary I/O devices.
We’re also told that AMD will stop using ASMedia to design its chipsets. But AMD says that it will continue to use ASMedia, but the new chipsets will be “more nuanced.” That could mean that AMD will design parts of the chipset and simply license key IP blocks, like USB 3.1, from ASMedia.
AMD has promised AM4 CPU socket compatibility until 2020 for all its Ryzen processors. That means you should be able to use any AMD Ryzen processor on any AM4 motherboard, providing AMD’s customers with a solid upgrade path in the future. That stands in stark contrast to Intel’s frequent socket changes that find enthusiasts having to migrate to new boards and chipsets. AMD’s long-lived support for the AM4 socket has earned plenty of cachet with enthusiasts, but it also restricts the company’s options for the new 3000-series processors.
The AM4 socket gives us a few hints about the possible configurations. The AM4 socket supports two channels of DDR4 memory, and that isn’t likely to change with the 3000-series models because each memory channel requires its own dedicated pins for communication. Given the current alignment of the AM4 socket, that means the new chips will likely have a dual-channel memory controller.
We also don’t expect a move forward to DDR5 in the near future, as we haven’t seen enough signs in the broader motherboard and memory module ecosystem to indicate there will be a wholesale move to the new standard next year. While its market share is growing, AMD doesn’t likely have the pull to drive the move to DDR5. That will likely have to come from Intel.
The 7nm Process
AMD hasn’t said much about the 7nm process that it will use for the Ryzen 3000-series chips, though the company claimed during its debut of the EPYC data center chips that the process will bring twice the density while halving the power consumption at the same performance level. The company claims 7nm also offers 1.25x the performance at the same power. These advantages will come to the end user in the form of faster and cheaper chips, but scaling difficulties and chip-level interconnect restrictions complicate matters as chips become smaller–see Intel’s ongoing issues with its own 10nm process.
AMD’s CTO Mark Papermaster tells us to expect a 25 percent increase in performance from the new process, but he reiterated to EETimes that the move to the new process is challenging:
“Moore’s Law is slowing down, semiconductor nodes are more expensive, and we’re not getting the frequency lift we used to get,” he said in a talk during the launch, calling the 7-nm migration “a rough lift that added masks, more resistance, and parasitics.”
Papermaster also said that the move to EUV (extreme ultraviolet) manufacturing, which will come with the 7nm+ node, will only provide “modest” device performance opportunities.
It is important to remember that node naming conventions have become more of a marketing exercise than a metric based on hard measurements. So TSMC’s 7nm is not denser than Intel’s pending, oft-delayed, 10nm node. In fact, Intel’s 10nm process is actually denser than TSMC’s 7nm process – Intel’s 10nm (6T) process is 100 mega-transistors per square millimeter (MTr/mm2), while TSMC’s 7nm (7.5T) weighs in at 66 MTr/mm2.
But for now, it’s a race to see who can get their new process to market first. Intel plans to have 10nm processors in high volume manufacturing in late 2019 (presuming there aren’t any more delays), leaving AMD a big window of opportunity if it can make its target of mid-2019.
The high up-front costs associated with developing a new node pushed AMD’s primary manufacturing partner, Global Foundries, out of the 7nm race last year. AMD remains committed to delivering the new node and using it as a vehicle to deliver the new Zen 2 microarchitecture to market, but it is partnering with TSMC for manufacturing. TSMC is the industry’s premiere third-party foundry, so AMD will have to compete for wafer output with big players such as Apple, Qualcomm and Nvidia, that also use the fab’s chip production facilities. However, recent reports indicate that the 7nm node is expensive, thus leading several large players to scale back product development on leading nodes, thus leaving about 10 percent of TSMC’s 7nm production capacity underutilized. This is a double-edged sword for AMD: while the company shouldn’t have any problem sourcing wafers from TSMC, the progressively higher costs of each smaller node means we might not see big price drops with the second-gen Ryzen chips.
The AdoredTV Leak
AdoredTV recently reported that an anonymous tipster had sent a list of AMD’s new processors. The leak claimed AMD will release the new Ryzen chips in Q1 2019, but several details of the leak are questionable.
The leaker contended that AMD would unveil the new processors and pricing at CES, but the company didn’t share such detailed information during its keynote. We expect the lack of detail about the Ryzen chips, as most of these final touches, especially CPU clock speeds and pricing, typically come later in the development process.
|CPU||Cores / Threads||GPU||Base / Boost Clock||TDP||Price||Debut|
|Ryzen 3 3300||6 / 12||–||3.2 / 4.0GHz||50W||$99||CES|
|Ryzen 3 3300X||6 / 12||–||3.5 / 4.3GHz||65W||$129||CES|
|Ryzen 3 3300G||6 / 12||Navi (15 CU)||3.0 / 3.8GHz||65W||$129||Q3 2019|
|Ryzen 5 3600||8 / 16||–||3.6 / 4.4GHz||65W||$178||CES|
|Ryzen 5 3600X||8 / 16||–||4.0 / 4.8GHz||95W||$229||CES|
|Ryzen 5 3600G||8 / 16||Navi (20 CU)||3.2 / 4.0GHz||95W||$199||Q3 2019|
|Ryzen 7 3700||12 / 24||–||3.8 / 4.6GHz||95W||$299||CES|
|Ryzen 7 3700X||12 / 24||–||4.2 / 5.0GHz||105W||$329||CES|
|Ryzen 9 3800X||16 / 32||–||3.9 / 4.7GHz||125W||$449||CES|
|Ryzen 9 3850X||16 / 32||–||4.3 / 5.1GHz||135W||$499||May 2019|
Several other facets of the leaked lineup are also questionable. The pricing model outlined in the leak would bring twice the number of cores to market within similar price ranges of the current AMD chips, which is highly unlikely given what we said above about the higher cost of 7nm production. AMD did turn the industry on its ear with its revolutionary pricing model with the first-gen Ryzen processors, but that release represented an entire reset of the company’s lineup.
Now the company is bringing an iterative update to market, meaning that such radical changes to its product stack would impact its existing lineup. Bringing 3000-series processors to market at these prices would cannibalize the company’s existing lineup, meaning it would have to purge its entire supply chain to avoid having unsellable excess inventory. That isn’t likely.
Others have pointed out more obvious holes in the lineup, such as TDPs and frequencies that don’t line up evenly across the product stack. We don’t think this leak is accurate, especially given the nature of the disclosure and the pricing model. But only time will tell if at least some of the leak contains accurate info.
For now, we wait
AMD’s move to producing 7nm chips in volume before Intel can counter with its own 10nm shrink could serve a huge blow to Team Blue, especially if AMD can also pull it off in the middle of this year.
But there are more gains to be had than “just” a node shrink: AMD also has its Zen 2 microarchitecture coming with the 3000-series Ryzen chips, and that should bring along a nice improvement in instructions per clock (IPC) throughput. That could potentially push the company closer to matching Intel in one of the few areas where the blue team still has a clear advantage–performance in single-threaded applications. AMD’s demo implies that it is reaching parity with Intel’s per-core performance, but we’ll need a much larger spread of workloads to accurately gauge the improvement.
AMD is already extremely competitive on many fronts. But if the company can close the single-threaded performance gap (which would be beneficial for gaming performance in particular), it could set the stage for the biggest upset in the history of the processor market. And that kind of success (paired with increasing EPYC sales) would give AMD a major influx of cash, which could help the company better compete against its much larger CPU rival (not to mention Nvidia on the graphics front) in 2020 and beyond.