Both the framework 16 and desktop use different versions of Ryzen AI chips I think.
But the desktop uses soldered ram and the laptop version uses normal ram. Is there an architecture or performance difference where these 2 chips operate very differently.
or would a 64 GB framework desktop be roughly equivalent to the 64GB laptop, for the AI chipset?
Both the AI 9 HX 370 and the AI MAX 395 run at 8000MT/s for soldered LPDDR5X. However the latter has much higher TDP, installing the latter to a laptop will seriously restrict its performance
So does that mean that for the current products.
Desktop is 8000 MT
laptop is 5000MT ? or so depending on what ram you use.
But otherwise mostly the same?
Not just the RAM, the CPU’s TDP. The AI MAX 395 can run at 120W
would a 64 GB framework desktop be roughly equivalent to the 64GB laptop, for the AI chipset?
If by ‘AI chipset’ you are referring to the NPU, then possibly but I’m not sure anyone cares … the NPU isn’t used by an awful lot right now and is very low powered compared to using the GPU.
As for the main processor itself, the AI MAX+ 395 is faster at multi-core performance than the HX370 and has 4 more cores (than the HX370) and when it comes to the integrated graphics, it is not even close with the AI MAX+ 395’s integrated 8060S GPU absolutely shaming the 890M in the HX370 chip.
https://www.cpu-monkey.com/en/compare_cpu-amd_ryzen_ai_max_plus_395-vs-amd_ryzen_ai_9_hx_370
The primary reason that the desktop uses soldered RAM but the laptop uses socketed RAM is down to that GPU. There’s no way to meaningfully feed that GPU performance with socketed RAM at this time.