Yeah you can't just plot this performance out of context. Ryzen chip performance is dependent upon the Ram, if there isn't enough or if there is the wrong type it will throttle and perform poorly.
I do not see how you can compare with this example this to me screams of a article full of bullshit. I do not even believe the article because of its attempt to skew the results not a fair game on any platform used. You as a writer need to fully understand the workings of architecture in these chips meaning use the same motherboard and ram for both chips. You are flawed on how to measure performance per W. Get real dude your not in nuclear now are you. You are here writing about s*** you have no clue works.
If you're not monitoring package power, you're not making a comparison.
See in example, the Zenbook Duo 14 new Tiger Lake model has a 25w? TDP compared to the regular Zenbook 14 Tiger Lake with a 19w TDP. And on this site as well as others, that difference was noted when comparing performance.
It would be like saying the 2021 g15 GA503 struggles against older CPUs such as found in the GT 76 Titan. This was a disappointing article.
This author does it purposely ... again. Comparing apples and oranges and using clickbait titles. Reading articles about AMD by Allen Ngo ...total waste of time.
Indeed, there are reviews on Youtube of the HP with 5800u running at a higher power envelope that performs much better on games. I expect the multicore scores would also be be significantly than the zenbook. Nowadays NBC seems to have difficulty drawing the obvious correlations 1) Cezanne offering higher single-thread than Luciene but the same core / clock specs 3) more power on these core U or core g processors means better multirhead/gaming performance.
The HP has 48W and 37W power limits whereas the Zenbook has 30W and 25W.... The comparison is not really far. You should set the power limits of the Zenbook on the HP and rerun the tests.
while the article is misleading, there are already reviews that compare 4800U with 5800U on similar laptops and chassis and they are quite close in MT workloads and GPU performance. Why is this? Because 5800U is just made on the same 7nm process and zen 3 doesn't have nearly enough efficiency improvements in terms of architectural enhancement that would let it increase the perf/watt metric by a considerable margin. The same situation holds for 5800H/4800H, where 5800H is a tiny bit faster than 4800H but it also consumes more power, hence the perf/watt is basically the same. Zen 2 brought such a huge efficiency increase over 3000 series because (and AMD said this in their slides) of 7nm. I recall they said it was about 70-75% of the improvements, which is a LOT. So zen 2 to zen 3 just shows what amd can do on the same node, which isn't a whole lot more, if they are TDP constrained.
It is true that the article is clickbaity. However, for most users, there is really no difference, Lucienne or Cezanne. Both come with LOTS of CPU cores/threads, that offer more capability than most people will ever need. And gaming performance is identical too, both at 15W and 25W (https://youtu.be/edpdmbmYOj0?t=837). And battery life in Lucienne is improved over Renoir (recent comparison here on Notebookcheck). So I'd say, if you need those Zen3 cores to do your work faster (that is- you are earning money with the device)- then sure, it is worth searching for Cezanne, and paying extra for it. And even better- just get H-series then, for higher sustained clocks. But for an average consumer- it is probably best to concentrate attention to other components (good screen, enough RAM, etc), instead of overpaying just to get somewhat faster Zen3 cores.