AMD
Both Intel and AMD often have processor updates to announce at CES in January, however AMD is not ready to introduce its next-generation flagship laptop chips: the Ryzen 8040 sequence is coming to laptops beginning in early 2024, although at first blush these chips look awfully just like the Ryzen 7040 processors that AMD introduced simply seven months in the past.
Though the generational branding is leaping from 7000 to 8000, the CPU and GPU of the Ryzen 8040 sequence are almost equivalent to the ones in the 7040 sequence. The chips AMD is asserting at this time use as much as eight Zen 4 CPU cores and RDNA 3-based built-in GPUs (both a Radeon 780M with 12 compute models, or Radeon 760M or 740M GPUs with 8 or 4 CUs). The chips are manufactured utilizing the similar 4 nm TSMC course of as the 7040 sequence.
There’s additionally an AI-accelerating neural processing unit (NPU) that AMD claims is about 1.4 occasions quicker than the one in the Ryzen 7040 sequence in giant language fashions like Llama 2 and ONNX imaginative and prescient fashions. Both NPUs are based mostly on the similar XDNA structure and have the similar quantity of processing {hardware}—AMD says that the AI efficiency enhancements come largely from increased clock speeds.
-
The preliminary run of Ryzen 8040 CPUs, with fashions designed for thin-and-light laptops and beefier workstations and gaming laptops.
AMD -
This slide from the authentic 7040 announcement exhibits simply how shut the 7040 and 8040 chips are, not less than by way of core rely, clock pace, and TDP.
AMD
Clock speeds throughout the board are fairly just like the 7040 sequence, although AMD Technical Marketing Manager Donny Woligroski claims that optimizations to firmware and drivers ought to nonetheless end in a “actual generational enchancment.” There’s loads of room to enhance a laptop chip’s efficiency with out growing peak clock speeds, particularly in the event you can enhance the sustained clock speeds that the chips can hit once they’re busy for an prolonged interval. But generation-over-generation efficiency comparisons for something apart from AI weren’t included in AMD’s presentation.
AMD’s briefing and slide deck spend most of their time speaking about Ryzen AI and the devoted NPU. Though Nvidia dominates the marketplace for server-side AI chips, native NPU {hardware} from AMD, Intel, and varied ARM chipmakers could make it simpler to run some generative AI workloads domestically. Depending on how these options are carried out, they might be higher for privateness since the processing is occurring on-device moderately than sending information to a server someplace—when Apple talks about privateness, it typically talks about this sort of on-device processing since its chips have included NPUs for years longer than Intel’s or AMD’s.
-
The solely direct generation-over-generation comparisons AMD made have been about AI efficiency; these come primarily from boosted NPU clocks.
AMD -
AMD in contrast Ryzen 8040 favorably to Intel’s Thirteenth-gen laptop CPUs, however did not present comparisons to any of its personal merchandise.
AMD -
Strix Point, coming later in 2024, could also be a bigger improve. All AMD is speaking about for now could be the next-generation XDNA 2 NPU structure, which the firm says is 3 times quicker than its present NPUs.
AMD
The fast leap in generational branding is enabled by the new processor naming scheme that it began utilizing late final yr. Rather than denoting something about the underlying {hardware}, the first digit of the four-digit mannequin quantity now simply tells you what yr the processor was launched. The second digit denotes the “market phase” (a 5 or 6 for Ryzen 5, a 7 for Ryzen 7, and so forth), and the third digit is the place you truly get technical details about the underlying CPU structure. The last digit is one other option to distinguish between lower- and higher-end chips, and a suffixed letter signifies the TDP vary of the chip and the sorts of techniques it may be utilized in.
Intel just lately criticized this naming scheme, declaring in a since-deleted slide deck that it permits AMD to rebrand older processors as newer ones whereas making minimal technical enhancements. This is totally honest criticism (and Intel quoted Ars whereas making it), although Intel repeatedly does the similar factor with its processor generations.
AMD
For now, this naming system solely applies to AMD’s laptop processors. Its Ryzen desktop CPUs, not less than to date, have solely jumped numbers when AMD has truly modified CPU architectures. AMD did not have a lot so as to add about a true next-generation successor to the 7040/8040 sequence silicon, however it did briefly tease a processor codenamed “Strix Point” that may launch in some unspecified time in the future in 2024 with a next-generation NPU based mostly on the XDNA 2 structure. AMD says Strix Point will function 3 times quicker generative AI efficiency than its present silicon.
Listing picture by AMD