Skip to main content

AMD Unveils 6nm Ryzen 6000 ‘Rembrandt’ Chips With Zen 3+, RDNA2 and DDR5

AMD
(Image credit: AMD)

AMD announced its Ryzen 6000 “Rembrandt” mobile processors at CES 2022 as it prepares to fend off Intel’s Alder Lake looming mobile processors and Apple’s exceedingly competitive M1-based systems. AMD has plenty of new additives to fuel the new chips, with headline features including the new Zen 3+ CPU architecture and RDNA 2 integrated graphics fabbed on TSMC’s 6nm node. The new Rembrandt chips also reach speeds up to 5.0 GHz. 

AMD says Rembrandt serves up 1.3X faster processing and 2X faster graphics than the 5000 series, and they’re also the first chips to support ray-tracing on integrated graphics. In addition, AMD touts improved connectivity across the board, including support for PCIe 4.0, LPDDR5, and DDR5, among other additives, all while delivering up to 24 hours of battery life. These new chips come to market in February. 

AMD also announced its new 3D V-Cache equipped Ryen 7 5800X3D and the coming Zen 4 Ryzen 7000 chips with 5nm tech for desktop PCs. Let’s dive into the details.

AMD Ryzen 6000 Series “Rembrandt” Mobile Processors 

(Image credit: AMD)

Here we can see the full stack of the Ryzen 6000 models that replace the Ryzen 5000 ‘Cezanne’ lineup. The H-series slots in for gamers and creators. AMD divides the lineup into the overclockable HX models with a 45W+ TDP for the fastest gaming laptops, 35W HS models for the ‘compact performance’ category, and the 45W H-series models.  

The four Ryzen 9 models come with eight CPU cores and 16 threads paired with 12 RDNA 2 CUs that boost up to 2.4 GHz. That’s a notable improvement over the Vega graphics units that came with the Cezanne processors, not to mention that Rembrandt comes with up to four more CUs than the Cezanne chips. The overclockable flagship Ryzen 9 6980HX leads the lineup with a 3.3 GHz CPU base, 5.0 GHz boost, and a 45W+ TDP. 

The Ryzen 7 lineup has eight cores and 16 threads paired with lower CPU clock rates and a lower peak GPU boost of 2.2 GHz, while the Ryzen 5 lineup comes with six cores and 12 threads that peak at 4.5 GHz paired with six RDNA 2 CUs that reach peaks speeds of 1.9 GHz. 

The U-Series appears to be a bit more straightforward, spanning from 15 to 28W for Ultrathins, but three of the models come with the aged Vega graphics engine that will result in significantly reduced iGPU performance.

AMD hasn’t shared the full details yet, but we do know that the Zen 3+ architecture is, as the name applies, an iterative update over the existing Zen 3 architecture that powers the Ryzen 5000 chips. 

We’re told there aren’t any changes at the microarchitectural level (i.e., the cores). However, the move from 7nm to 6nm allows the company to cram in more features per square millimeter of silicon — TSMC’s 6nm process offers up to 18% higher logic density than the 7nm “N7” process, but similar power and performance metrics. The 6nm “N6” process also works with the same design tools as N7, easing the porting process. In addition, N6 uses EUV lithography for up to five layers of the chip, thus reducing costs and speeding production. 

ProcessorProcessDensity
Ryzen 4000 -Renoir7nm156 mm²
Ryzen 5000 - Cezanne7nm180 mm²
Ryzen 6000 - Rembrandt 6nm208 mm²
Alder Lake -PIntel 7217 mm²

Here we can see that, despite the move to the denser 6nm node, the Rembrandt die has grown by roughly 16%, which is obviously needed to accommodate the more expansive feature set, like four more CUs than the prior-gen models. 

Interestingly, AMD cites better performance/watt with the 6nm process, but that doesn’t mesh with TSMC’s projections of similar power/performance. In either case, Ryzen 6000’s performance and battery life improvements also stem from the 50 new and enhanced power management features, adaptive power management framework, and new deep sleep states. 

AMD’s RDNA 2 iGPU is the first ray tracing-capable APU on the market. Rembrandt features up to four more CUs than Vega, topping out at twelve RDNA 2 CUs that can boost up to 2.4 GHz. Rembrandt comes with a 1.5x larger GPU compute engine, 1.5x more memory bandwidth, and 2x larger L2 caches and render backends, all of which are sorely needed improvements over the aging Vega architecture in the previous-gen APUs. 

It’s also nice to see the RDNA 2 graphics engine fed by LPDDR5/DDR5, as the enhanced memory throughput should greatly improve performance. Rembrandt also steps up to PCIe 4.0, and we imagine that applies to the connection to the GPU, too.

AMD says its new power tech and the 6nm process culminate in 24-hour battery life and shared some of its internal benchmarks. As always, take vendor-provided benchmarks with a grain of salt, but this applies more in this case than usual: AMD dialed in a 28W power setting for its 6000 series chip but left the previous-gen model at 18W. That means these benchmarks are misleading.

The company claims Ryzen 6000 has 30% lower power consumption during video conferencing, reduces power by 15% during web browsing, and delivers a 40% savings during video streaming (all relative to its previous-gen 5000 series). 

It’s easy to have lower power consumption if you simply dial down performance, but AMD’s application benchmarks claim that these power savings come in tandem with a big leap in performance. Compared to the Ryzen 5800U, AMD claims that the Ryzen 7 6800U offers up to 1.3x the multi-thread performance, 1.1x the single-thread performance, and 1.3x more performance in productivity applications. Looking at more intensive workloads, AMD claims that encoding, rendering, and gaming improve by 1.7x, 2.3x, and 2x, respectively. 

This is where the rubber meets the road for another set of misleading benchmarks. Again, AMD used a 28W power setting for its newer chips, but left the older model at 18W.  Here we can see the Ryzen 7 6800U take on the Intel Core i7-1165G7’s integrated graphics and the entry-level MX450 discrete GPU in a series of gaming benchmarks, with all of the wins going in Rembrandt’s favor. AMD also highlighted Rembrandt’s generational uplift in a series of games, with the 6800U taking a 2x lead over the 5800U in nearly every title tested.

AMD also announced at CES that Far Cry 6 now supports FidelityFX Super Resolution (FSR) and ran a demo that hit 59 fps on stage with the quality preset, a 51% improvement over playing with the feature disabled. As shown on the last slide, expanding out to a broader list of titles like Deathloop, Godfall, and Call of Duty: Vanguard implies that FSR could be the killer app for mobile gaming. 

AMD has made big strides on its platform connectivity tech, adding support for LPDDR5 and DDR5 memory, PCIe 4.0 connectivity, USB4 up to 40 Gbps, WiFi 6E with dual-band simultaneous capabilities, Bluetooth LE 5.2, and AV1 hardware decode (unfortunately, we don’t see AV1 hardware encode support). AMD threw in AI noise cancellation tech for good measure, but we don’t know any details yet. 

Rembrandt is also the first CPU to support Microsoft's Pluton technology. Pluton enables more robust security that helps prevent physical attacks and encryption key theft while protecting against firmware attacks, a welcome feature when you're on the go with your laptop. This tech originally debuted in the Xbox and AMD's EPYC data center processors and joins AMD’s other security features, like AMD Secure Processor and Memory Guard, among others. 

All told, the company expects to have 200+ premium systems in the market in 2022, but that includes the entire Ryzen portfolio (5000 series, too). AMD plans for the first Ryzen 6000 mobile systems to start shipping in February. AMD has also tremendously improved the turnaround time required to bring its mobile APUs to desktop PCs. We can’t wait to see how Rembrandt performs outside of the thermal and electrical limitations of mobile devices.

Paul Alcorn
Deputy Managing Editor

Paul Alcorn is the Deputy Managing Editor for Tom's Hardware US. He writes news and reviews on CPUs, storage and enterprise hardware.

  • brandonjclark
    Ray-tracing on integrated graphics?


    Noooooo!
    Reply
  • tommo1982
    In was hoping Rembrant will come to desktop as well. I'd change my Picasso then. Tough luck for us desktop users, again.
    Reply
  • hannibal
    Laptop market is much, much more important than desktop market!
    If the chip shortage end by the end of 2024 as TSMC predict, maybe then also desktop versions...
    But in anyway this tells well why "low end" gpus are essentially dead...
    Reply
  • linuxdude
    Any idea why in the table the chip area is called density?
    Density is transistors/mm2, mm2 is simply area.
    Reply