Leaker thinks Steam Deck 2 will "run circles" around current handheld consoles

[Redaktion]

The AMD Gainsborough APU that could power the Valve Steam Deck 2 recently leaked online. AMD Gainsborough's potentially small iGPU with its low TDP doesn't inspire much confidence in its ability to run AAA games. However, Moore's Law Is Dead thinks people are underestimating the performance of the Steam Deck 2.

https://www.notebookcheck.net/Leaker-thinks-Steam-Deck-2-will-run-circles-around-current-handheld-consoles.1420066.0.html

[+ ~2x FPS/Watt]

This shows that RDNA 4 has a huge advantage over RDNA 3/3.5.

RDNA4 = RDNA3/3.5 + proper hardware upscaling instructions (FP8), and RDNA4 is 7-11% more energy efficient vs a 7900XTX (most efficient RDNA3).

And RDNA4m could be exactly that: RDNA3 plus hardware upscaling instructions (FP8) minus everything else that is not required for a 10-15W APU.


AMD Ryzen Z1 APU is using TSMC's 4N node (slightly improved 5N -- basically 5N), and by 2028 TSMC 2N node will be available in the sense that it's also going to be cheap enough by then, then using 2N node, the raw performance would have improved by 2 nodes (5N -> 3N -> 2N).

Using a full 2-node improvement: 1.15 * 1.15 (= 1.15^2) = +32% FPS per Watt improvement (it used to be more than 15% per full node, btw.).

Using a 3-node improvement / including TSMC A14 node:
1.15^3= +52% FPS per Watt improvement.

Only +50% FPS per Watt even after 3 full node shrinks shows that proper (=FP8, not INT8) hardware upscaling capability will be a priority.

Using hardware-based upscaling (FP8+FSR4, or better) on top of the nodes improvement is going to to certainly improve the perf by 2x (2 = 1.52 * 1.3 (what upscaling at good quality delivers)).