Advanced Micro Devices Inc., Intel Corp. and Nvidia Corp. all updated their respective product portfolios at CES 2022 today by introducing new chips for the consumer market.

Intel’s Mobileye self-driving vehicle division also debuted its latest silicon. Mobileye detailed an upcoming chip called the EyeQ Ultra that targets autonomous driving use cases and offers 10 times the performance of the division’s earlier EyeQ5 processor.

Laptops in focus

AMD introduced a new line of central processing units for laptops today that will be marketed as the Ryzen 6000 Series. The company said customers can expect 11% faster single-threaded performance compared with the previous-generation Ryzen 5000 Series. For multithreaded applications, the company is promising a performance boost of up to 28%.

The CPUs’ increased speed is the result of several factors. One contributor to the performance improvement is that the chips are produced using a six-nanometer manufacturing process, an upgrade from the seven-nanometer technology that the previous Ryzen 5000 Series uses. 

At the architectural level, the Ryzen 6000 Series is based on an enhanced core design dubbed Zen 3+. AMD said that Zen 3+ includes power management features that allow it to adjust speed levels more efficiently than earlier designs. 

Each of AMD’s new laptop chips combines multiple Zen 3+ CPU cores with an integrated graphics processing unit. The integrated GPU is based on the company’s latest RDNA 2 graphics architecture. According to AMD, customers can expect an up to twofold increase in GPU performance compared with previous-generation silicon.

The integrated graphics cards in the Ryzen 6000 Series feature as many as 12 compute cores, a 50% increase from before. Moreover, the cores can operate at a 20% higher clock frequency. For an additional performance increase, AMD doubled the onboard cache that the graphics cards use to store data for processing.

Each of the 10 CPUs in the Ryzen 6000 Series offers a different mix of performance and efficiency. The entry-level Ryzen 3 5425U features four cores that run at a top frequency of 4.1GHz. The flagship Ryzen 9 6980HX, meanwhile, has twice as many cores and a maximum frequency of 5GHz.

Increased competition

Intel also doubled down on the laptop market at CES by introducing no fewer than 28 mobile processors. The lineup is headlined by the company’s new flagship laptop CPU, the Core i9-12900HK, which it claims is the fastest mobile processor ever made.

The Core i9-12900HK can run certain creative applications up to 44% faster than its predecessor, Intel claims. Processing speeds are 28% higher for certain videogames. The Core i9-12900HK’s performance is provided by 14 onboard cores that Intel said can reach a maximum frequency of 5GHz.

Intel has decided to incorporate not one core design but two into its new flagship laptop CPU. Six of the processor’s cores are based on a performance-optimized design built to maximize processing speeds. The remaining eight are based on an efficiency-focused design, which trades off some speed to run applications using less electricity. 

Like AMD, Intel focused focused primarily on the consumer market with its CES 2022 product announcements. But the company also took the opportunity to provide an update on the work of its Mobileye division. Mobileye, which Intel bought for $15.3 billion in 2017, is a major supplier of processors and other components to the auto sector.

The Intel subsidiary debuted a five-nanometer chip dubbed the EyeQ Ultra today intended to power Level 4 vehicles. A Level 4 vehicle is capable of driving itself without any human input in most situations. The EyeQ Ultra can carry out 176 trillion computing operations per second, Intel says, which allows the processor to provide performance equivalent to 10 of Mobileye’s earlier EyeQ5 vehicle chips.

Mobileye expects to ship the first EyeQ Ultra units to customers in late 2023. Mass production is expected to begin in 2025.

New flagship GPU

Alongside AMD and Intel, Nvidia shared its own series of product updates at CES 2022. The company introduced improvements to its Omniverse simulation and design software along with a new flagship consumer GPU, the GeForce RTX 3090 Ti.

The GeForce RTX 3090 Ti provides 40 teraflops of performance for running shaders, specialized programs that GPUs use for graphics rendering. A teraflop equals a trillion calculations per second. The new GPU also provides 78 teraflops for ray tracing, a technique that applications use to render realistic lightning effects, and Nvidia said that the chip also provides a hefty 320 teraflops’ worth of computing capacity for machine learning workloads.

Image: Intel