Although this year’s CES was far from ordinary with the pandemic still near and protocols still in place, it was a very exciting CES. I have already written about Qualcomm’s achievements in the realm of automotive and ADAS systems. I have also already written on the top notebook trends at CES 2022, of which these new technologies will go into. Now I want to turn my attention to Intel, NVIDIA, and AMD processor innovations that were announced at CES 2022. All three have announced new processors that I believe should make for a competitive and exciting year.
Many new products were announced by Intel AMD and NVIDIA that cover next-generation CPUs, next-generation and all-new GPUs, and next-generation automotive processors. In each category, I believe there is a lot of innovation and competitiveness. I talked about Intel and AMD at CES on the Six-Five podcast. Let’s take a look at what Intel, NVIDIA, and AMD announced at CES 2022.
When talking about CPUs and GPUs, we have to consider the different audiences for these processors. There are about four different audiences that make up the processor market including business, consumer, gaming, and content creation, with some of these overlapping. The emphasis on mobility and PPW was one trend that stuck out to me for next-generation processors. I also noticed the trend of GPUs allowing for AAA and 1080p gaming to be more accessible on a mobile platform. In the Automotive space, Intel Mobileye and NVIDIA announced dates for AD technologies and they seem much more real.
Intel notebook processors
Interestingly enough, Intel did attend CES 2022 physically, recording its online announcement from the Mandalay in the same space the live and in-person press conference was supposed to be. Intel had three large announcements that I will break down into the categories of CPU, GPU, and automotive.
Intel released its 12th Gen desktop processors last year at the end of the year, which was a big deal given its new asymmetric architecture. Intel’s 12th Gen adopted an iteration of the BIG.little-style architecture with performance and efficiency cores for managing different processes. Intel is released 22 new 12th Gen processors across its entire processor family.
In previous generations, Intel has had H-series and U-series processors for notebooks. The H-series is the beefier iteration used in larger mobile workstations and gaming notebooks. H-series laptops have really taken on a different slimmer form. Intel’s keynote at CES showed base power as a measurement of power rather than TDP. Intel reasons that TDP does not effectively show the power and battery draw on hybrid cores as it does with legacy cores. I say this pointing towards Intel’s Power-Performance graph of last generation’s Intel HK-series processors, Ryzen’s last generation HX-series processor, and this year’s Apple M1 Max processor against Intel’s newest HK-series mobile processor.
The important aspect to look at in the 12th Gen’s benchmarks is its answer to Apple M1 Max’s PPW. The PPW of the M1 Max is impressive, to say the least, and from my owning testing, the PPW of the M1 Max outpaced Intel’s 11th gen H-series but not in raw performance. See my blog on the benchmarks here. I like Intel’s graph because it takes data from Apple’s own graph from its original M1 Max keynote back in October. It then takes 11th Gen benchmarks, which are similar to my benchmarks on DC, and compares the two with its 12th Gen equivalent. Not only does it tell us that the 12th Gen is capable of 40% more performance over the last generation, but it also tells us that it has a better PPW over the competition. The H-series has a base power of 45W and a Turbo power of 115W for the i9 and i7 models and 95W for the i5 models. What I found interesting is that the H-series has 8 E-cores and 6 P-cores, two more E-cores than P-cores. I believe Intel used this ratio because it is the best ratio for performance while regulating thermals and power draw. It leads me to believe that Intel is putting a lot of confidence in its Thread Director to manage each core’s compute. What is also interesting to me about the P-cores and E-cores is that each P-core has two threads, and each E-core has one thread. I imagine if Intel configured the E-cores to take on extra threads, the processor would be much closer to the turbo power of 115W than the base power of 45W, resulting in less efficiency.
Intel also announced a new P-series that sits between the legacy H-series and the U-series for ultra-mobile processing. In previous years, Intel had only the H-series for mobile workstations and the G7-series for its thin and light series. The hybrid architecture opens up the possibility for two types of mobile processors, the P-series, and the U-series.
The P-series matches the H-series in base power with 45W across the board and 65W Turbo Power across the board. The biggest difference between the P-series and the H-series is that it has fewer P-cores compared to its H-series equivalent. I see the P-series essentially as an H-series that has been slimmed down to meet the requirements of the Intel Evo Platform. Fewer P-cores and a lower power draw allow Intel to put really power cores in a smaller form factor while maintaining the performance that up to 6 P-cores can bring. I believe it also puts less pressure on the U-series to maintain a performance level that is expected of ultrabooks and to focus more on the ultra-mobile form factor.
Intel’s U-series processors again bump down the base power and the number of P-cores for extra-mobile processors. Intel has two sets of U-series processors that can either have a 15W base power and a 55W Turbo Power or a 9W base power and 29W Turbo Power. With the new hybrid architecture, Intel is also putting more cores per processor with the i7-1265U 15W and 9W having a total of 10 cores and 12 threads. To put into perspective the upgrade, last year’s 11th Gen G7-series capped out at 4 cores.
While I haven’t directly compared AMD’s new offerings to Intel’s yet, my instinct tells me from the data I have seen is that Intel moved the net needle forward.
Intel discrete notebook graphics
Intel also announced its new high-performance Arc discrete graphics for notebooks. Intel says it will offer advanced features such as hardware-accelerated Ray Tracing, X^e Super SamplingAI-driven upscaling technology, and Intel Deep Link technology. Intel Deep Link is its new processing technology that allows for a system to utilize discrete and integrated graphics to share a workload. By sharing the workload, processes like rendering a 3D model can be quicker than simply utilizing one GPU. This technology isn’t new as AMD and Radeon had this kind of feature, but a decade a ago. AMD’s implementation didn’t work great if the tweo GPUs were too asymmetrical. I look forward to the tests.
Intel has already partnered with many OEMs to implement Intel Arc into systems like Alienware. I saw this as a big deal as Alienware is always the hardest gaming notebook to get into, but then again, Dell Tech owns Alienware that is all Intel all the time. I believe competition is always good, and I think it is good that Intel is getting serious with its GPUs; no longer just integrated graphics. What’s important to see is that Intel’s roadmap for discrete GPUs tells me Intel is serious about its high-performance GPUs.
Intel also announced its next-generation Mobileye SoC for autonomous vehicles (AV) called EyeQ Ultra. I have written on Qualcomm’s automotive plays at CES, including its new ADAS and new partnerships that will be realized as early as 2023. While there are many moves towards advanced driving assistance systems (ADAS), which in terms of implementation is a step before AV, Intel is looking ahead towards L4 with its EyeQ Ultra. Intel says its new SoC is capable of handling all the needs of L4 applications. It was engineered in tandem with Mobileye software and packed the performance of 10 EyeQ5s with 176 tera operations per second (TOPS). Intel also announced the EyeQ6L, which succeeds the EyeQ4 for premium ADAS at 55% of the EyeQ4’s size. I believe Intel is making leaps and bounds with its automotive technologies while simultaneously considering how this technology will be applied.
In the Mobileye keynote by Prof. Amnon Shashua, Mobileye CEO, he broke down the two directions in which AV could go. The first is towards robotaxis and then consumer AV. However, the unfolding of both directions has to be consumer AV to robotaxi rather than robotaxi to consumer AV. The reality is that the consumer AV direction has the ability to add robotaxi layers on top and starting with robotaxis would start geo-fenced. Since the implementation of ADAS, which has a direct path towards AV, comes before consumer AV and robotaxis, the unfolding of the autonomous mobility market should unfold with consumer AV to robotaxis. The reason is that robotaxis are only in the L4 stage of the autonomous mobility market. Intel’s roadmap shows that L4 Robotaxis are coming as early as 2022 and 2023. Still, unless robotaxis become a majority at scale by 2024/2025 (according to Intel’s roadmap for Consumer AV), the consumer AV should follow ADAS before robotaxis do. In other words, I believe robotaxis are on a completely different timeline than consumer AVs because of the necessary technologies that need to be in place and the ability to scale both platforms.
Mobileye’s ADAS business is 4X larger than NVIDIA’s and needs to be taken seriously. I will be watching the details of the upcoming spin carefully.
Similarly, NVIDIA has announcements for new GPUs.
NVIDIA announced the adoption of its DRIVE Hyperion platform, including electric vehicle makers Polestar, NIO, Xpeng, LI Auto, and IM Motors for consumer AV. From the robotaxi side, Cruise, Zoox, and DiDi also embrace DRIVE Hyperion and for trucking services Volvo, Plus, and TuSimple, who made an announcement at CES, are also embracing also looking to use NVIDIA’s DRIVE Hyperion. I believe NVIDIA is taking a similar route to Intel Mobileye in considering the adoption of its high-performance computer and sensor architecture for consumer AV and commercial AV (robotaxis and trucking services). I believe NVIDIA addresses the unfolding of both categories of AV with its intelligent assistance.
NVIDIA showed a demo of its DRIVE Concierge and DRIVE Chauffeur intelligent assistant. The DRIVE Concierge takes the technology of Omniverse Avatar to be the driver’s assistant while driving, providing helpful information and even booking reservations. DRIVE Chauffeur, on the other hand, is an AI-assisted driving platform that is based on NVIDIA’s DRIVE AV SDK. NVIDIA says the DRIVE Chauffeur pipeline has been architected to perceive 4D for robust and efficient autonomous operation. The Drive Concierge includes codes to train AI models in the cloud and run complex simulations in data centers for testing and validating AV systems. I believe this allows the consumer AV and commercial AV categories to be much more similar in terms of their respective rollout.
Transitioning back to GPUs, NVIDIA also announced new notebook discrete graphics cards, including the GeForce RTX 3080 Ti with 16GB of dedicated memory and the GeForce RTX 3070 Ti. NVIDIA includes its 4th generation of Max-Q technologies, focusing on GPU/CPU power use and battery discharge. NVIDIA says laptops are running Max-Q provide 70% more battery life with the extra performance. That extra 70% of performance comes at the perfect timing when the newest CPUs are also optimizing for the best PPW. Out of the 160 thin and light laptops using RTX 30 Series GPUs, I am curious whether or not any of them are on Intel’s Evo Platform. I believe it won’t be long before we see that type of combination in high-performance CPU and GPU for thin and light, battery-oriented notebooks.
AMD notebook processors
AMD announced new Ryzen 6000 series processors for laptops built on its “6nm” Zen 3+ core architecture and includes AMD RDNA 2 architecture for integrated graphics. It is the first time the RDNA 2 architecture is in an AMD mobile platform, which allows for gaming on ultra-mobile systems. It will bring FSR to the Ryzen 6000 processors. AMD has had a record market share in the past couple of years, and in order to hold that market share within the laptops space, the Ryzen 6000 series mobile processors will need to match Intel not only in innovation, but in designed sockets and marketing spend. While the desktop processors market is a little more sticky since there are other components, like a motherboard, that is considered when switching between red and blue. The laptop market is not like that. For that reason, the laptop market is slightly more telling of technology market performance. When comparing the new 6000 series mobile processors, the measurement to look at is PPW.
The Ryzen 6000 series processors run at up to 5Ghz and the company says it provides 1.3 times faster processing speeds and 2.1 times faster graphics than the previous generation. AMD says that the new Ryzen 6000 mobile series have up to 69% better video encoding performance, 125% better 3D rendering performance, and 100% better gaming performance over the 5000 series mobile processors. The Zen 3+ architecture also has new adaptive power management features and a new adaptive power control framework. AMD showed power consumption numbers of 30% less power on video conferencing, 15% on web browsing, and 40% on video streaming. I am used to calling out companies for cherry-picked performance numbers. I do believe AMD did an injustice for not showing the power consumption for the category of gaming—the one “hat” AMD consistently wears and shows off.
I am impressed with the performance numbers over last generation’s 5000 series. Where AMD should achieve better PPW is in its smaller Zen 3+ architecture. I think the extra-mobile world we live in could emphasize mobility and PPW. Although it enables new power management features and a new adaptive power control framework, I don’t believe these features will have more of an effect on battery life than its architecture. Seeing as the new architecture is a jump to a “+” and not a “4”, I believe it is an indicator that AMD is trying to squeeze out as much performance before moving on to the next generation architecture. The unfortunate angle of that squeezing could be an “L” in PPW.
AMD desktop processors
AMD announced a new desktop processor with its 3D V-Cache technology, the Ryzen 5800X3D. The Ryzen 7 5800X3D has Zen 3 architecture and an AM4 socket, and according to AMD is the world’s fastest gaming processor, beating out the Intel Core i9-12900K. In AMD’s footnotes on the benchmarking, the Ryzen 7 5800X3D with DDR4-3600 memory showed a 5% performance difference over the Intel Core i9-12900K with DDR5-5200 memory. AMD then previewed its Ryzen 7000 series processors with “5nm” Zen 4 architecture and a new AM5 socket that should be available in the second half of 2022. AMD preview and new 3D V-Cached technology continue to get me excited for the future and the competitiveness of high-performance computing. Even though Intel is biting back after AMD has has ite record years, AMD is not giving up, and we can all appreciate that.
AMD desktop discrete GPUs
AMD also announced two new desktop graphics cards, the Radeon RX 6500 XT and the RX 6400 graphics for 1080p gaming. These graphics cards are on AMD’s RDNA 2 gaming architecture and should be available in Q1 of 2022. It will also bring along FSRat the budget level. AMD is not only making 1080p and AAA gaming more accessible, but it is also making its FSR technology available to a broader audience.
I believe we saw a lot of great desktop, notebook and auto innovation from Intel, NVIDIA, and AMD at CES 2022, and as the year rolls on, I am excited to see these products and services in action.
Intel and NVIDIA are putting dates on AD technologies that will be here before we know it. Considering how impactful AD technologies will be when most of our roads are filled with safe AVs, we should see fewer car crashes and deaths on the road. I believe that is the goal we are all looking towards, and EyeQ Ultra and NVIDIA’s DRIVE Hyperion platform have the performance to get us there. Until then ADAS with L2+ and L3 technologies will have to do and do make us safer.
I’m also excited to see Intel’s new 12th Gen mobile processors shake up the laptop market as it regains its competitive position. Out of the other side of my mouth I will remind people that Intel has 80% market share. AMD’s Ryzen 6000 series likewise shows promising performance gains, and I am excited to see Intel’s 12th Gen go head to head against AMD’s Ryzen 6000 series. While I haven’t directly compared AMD’s new offerings to Intel’s yet, my instinct tells me from the data I have seen is that Intel moved the net needle forward.
As all three companies ramp up their respective discrete GPU offerings, we should see great devices for content creation, hybrid workflows, and great gaming.
This year should be a good year to buy a new computer!
Note: Moor Insights & Strategy co-op Jacob Freyman contributed to this article.