Or display it wrong or it means you are not using the GPU.

With
5800H i have 11-17%
RTX3060 at 30-40%
I had a lot more stuff opened, >4000 handles and about 260 poocesses.

the Youtube file was 4K but my screen is FHD so there was some calculation to reduce the size.

Nice analysis! I agree with every aspect of it. Especially using battery power for the initial 64W turbo boost while in battery mode. And yeah, even with a power supply the PL2 phases tax the battery a bit.. This is just bad. They should've given us a 90 or 100 W brick and also an over-boost mode (45 W PL1) for work, which requires a lot of power, and silence isn't the priority.

I can't stand the glossy display and the bad low-travel keyboard (to me). A typical consumer product in which they put components, that seem "premium" at first glance (for the masses) and gives it a "premium look" instead of being practical. Wow, the Touchpad is a real mess on that thing.

Ah, but for the Vega iGPUs, we measure significantly lower loads on 4K YouTube playback. Like 1 % - see attached screenshot of the ProBook 445 G8. Does the task manager display it wrong?

YT4K60_Wiedergabe.jpg

Quote from: NikoB on September 17, 2022, 21:12:52[...]

Again, the problems are clearly visible. The weak PSU forced the engineers to limit the PL2 to 52W, although Intel claims 64W in boost. So the processor is NOT fast enough in pulse as it could be with a 95-100W PSU. At the same time, the engineers overclocked 1280p in PL1 for some reason, exceeding the 28W limit set by Intel, which only increases noise without much help in performance.

The second obvious drawback is the OLED screen. He is neither.
It does not have infinite black, and at the same time (as ASUS failed many times before), it still has low-frequency PWM. You can forget about Static HDR with such a screen, and DV (as an analogue of Dynamic HDR10 +) is again not supported. And as usual, there are two key drawbacks in comparison with IPS - a lower color resolution than stated by the manufacturer and a glaring glossy screen (for some reason they never make matte AMOLEDs). Well, as a stake in the "heart" of this panel, you can note that it was not possible to calibrate the screen to dE <2, which, by the way, is quite a characteristic phenomenon for AMOLED screens in laptops (and smartphones). IPS almost always wins in terms of color accuracy.

Well, I draw attention to the author's rather lengthy reasoning that the battery wears out in boost, without explaining in which case - with or without a power supply? If with a power supply unit, then this increases the wear many times over. But the author mentioned that there is nothing left of the luxurious performance even in the boost when running on battery power, the performance drops by 2 times. After all, the whole point of using such laptops is battery operation, and not from the PSU, right? Otherwise, why is it needed at all?

Quote"In intelligent cooling mode, PL2 is even set to 64 watts but only maintained for 15 seconds in that case."

This is not physically possible from a 65W PSU without using battery power in parallel when running from the PSU. Again, we are dealing with increased and rapid battery wear to help with insufficient PSU power? Is it normal? For me, this is unacceptable, without an explicit warning in the specifications of potential buyers.

Quote"We measured a pretty low CPU utilization of 12%. During 4K YouTube playback, however, the iGPU's utilization is high with almost 50%, which is typical for Intel."

This is typical for AMD as well. =) In general, against the background of smartphone SoCs that consume a few watts to play 4k video, it's just a shame for the entire x86 industry in 2022 to consume more than 5W in total to play 4k video on YouTube. The engineers of both Intel and AMD should have done hara-kiri long ago for such a technical disgrace, if they have the concept of honor ..

Quote"We measure prolonged delays of up to 2,532 µs on the Yoga Slim 9i, mainly caused by Windows' ACPI driver. It can lead to issues with real-time audio and video work."

Again, a total shame for the entire x86 industry - not a single laptop manufacturer in a compartment with M$ programmers,
Can NOT solve the problem of working with real-time applications. How long can sound engineers and just fans of perfect sound reproduction on external equipment wait for this shame to end? Of course, most of the blame here lies with the driver manufacturers, but the main one, of course, is with Microsoft, which does not give a damn about this critical topic.

Incidentally, Windows XP was the last real-time OS in terms of minimal system latency. I immediately discovered this by switching to W7 once. W10 naturally became even worse. No matter what I do with drivers and settings, I couldn't achieve the same ideal performance as under XP, where almost no cores load interrupts the reference sound reproduction. How it possible -  Microsoft and hardware manufacturers? Is this really progress from 2001 to 2022?

Well, the picture is completed by the fact that at a price of almost $ 2100, Lenovo supplied ridiculous 16GB (free 12-13 under W10 / 11) without the possibility of expansion. Only very unpretentious and narrow-minded people will buy such a shameful config. The minimum, once again, should be 32GB. Yes, and a 1TB SSD at this price is simply ridiculous, given how SSD prices are falling. 2-4TB would be much more adequate for such a price tag.

The only undoubted advantage of this laptop in current version bios and hardware -
almost reference tuning ram settings (which immediately had a super-positive effect on the AIDA64 Photoworxx test - Photoshop productivity and etc.), well, except for the increased latency. Why does Lenovo constantly screw up in other models, especially "gaming" ones, where memory speed is critically important in general? And this is especially important in all models without a discrete video chip. After all, the built-in video chip uses the same memory as the system ...

Many other laptop models from Lenovo has the same issues with touchpad. Did they forget to test it?