AMD has brought 4 nm Hawk Point and Phoenix APUs to business users in the form of Ryzen Pro 8040 series for laptops and Ryzen Pro 8000 series for desktops.
Like its predecessor, Ryzen Pro, APU is loaded with business-focused features, primarily superior security options and remote management. Newer generations of Ryzen Pro typically come with the latest enterprise features and new hardware, but that's not the case with the 8040 or 8000 series.
The Hawk Point architecture within the Ryzen Pro 8040 series is actually almost identical to the traditional Phoenix arch that powered the Ryzen Pro 7040 series. It uses the same TSMC 4nm process, and nothing has changed in terms of core count or frequency for the CPU cores and integrated GPU. The only notable difference is the faster XDNA Neural Processing Unit (NPU), which can reach 16 TOPS instead of just 10 TOPS on the original Phoenix chip.
The situation is a little different for the desktop Ryzen Pro 8000 chips. The previous generation of his Ryzen Pro 7000 desktop processors did not use Phoenix silicon, instead using AMD's CPU chiplets found in high-performance desktop, workstation, and server CPUs. In terms of raw CPU core performance, the top-end Ryzen 7 Pro 8700G with 8 cores can be significantly slower than the 12-core Ryzen 9 Pro 7945.
AMD is clearly hoping Ryzen Pro 8000's improved efficiency, significantly faster integrated graphics, and dedicated NPU will make up for it. We can only assume that the larger Radeon iGPU and NPU combined can reach 39 TOPS, and the Ryzen 9 Pro 7945 with a smaller iGPU and no NPU can only achieve a few TOPS at best.
Not all Ryzen Pro 8040 and 8000 processors come with XDNA NPUs. The lowest-end 6-core and 4-core APUs are powered by a completely different chip called Phoenix 2. It uses two regular Zen 4 cores and four Zen 4c cores, with higher density but lower clock speeds. The Phoenix 2 just doesn't include an NPU, and the Radeon 740M graphics are noticeably slower than the regular Phoenix-powered graphics, so these Phoenix 2 APUs will be significantly slower on AI workloads.
AMD really wants people to use AI for business.
AMD is doubling down on AI with its latest Ryzen Pro chips, especially since improved AI performance is the only selling point compared to previous models. The company says AI can save time by doing tasks that would otherwise be done manually, such as writing documents or emails based on “a few bullet points.” However, you need to trust the AI to get it right and not waste meeting minutes trying to understand its word salad.
It also suggests using AI assistants for coding, but there's good reason to think that not all developers will find that an attractive feature.
If any company wants to use CPUs with AI, AMD says they should use its own chips instead of Intel's. His Hawk Point-based 8040 APU could reach a total of 39 TOPS, which is considerably higher than the 34 TOPS that Intel claims its Core Ultra CPUs with Meteor Lake silicon can perform.
On desktops, the gap is even wider as Intel's desktop Meteor Lake chips are being discontinued and it relies on the 14th generation Raptor Lake Refresh model. This model does not have his NPU at all and has rather weak graphics like the older Ryzen Pro 7000 series.
Not surprisingly, PC companies like HP and Lenovo, like Microsoft, say AMD's latest AI-powered desktop and laptop processors are superior. HP already has a strong lineup of PCs with Ryzen Pro 8040 and 8000 chips, while Lenovo has launched several ThinkPads with his Ryzen Pro 8040 APUs and ThinkCentre desktops with Ryzen Pro 8000 chips. Did. ®
