Easa

Registry entries added. After running the command: "Vypnuté" means OFF. WHEA Errors still present, they were there even before I had the 990 Pro drives inside. Nothing changed, yet. This: https://www.dell.com/support/kbdoc/en-us/000216115/laptops-with-12th-gen-and-12th-gen-hx-intel-core-processors-may-display-warning-message-whea-loggerid17 eh? 😄

powercfg /qh SCHEME_CURRENT SUB_PCIEXPRESS PCIEXPRESS_ASPM_STATE throws invalid parameter I have done a powercfg /energy report, and it says that the PCIE ASPM is disabled when running on AC Power. What I am also observing that the machine sometimes freezes for like 2 seconds with no apparent reason, sometimes its when maximizing window, sometimes its when launching an app or opening context menu. No clear link to anything.

No unexpected shutdowns No WHEA errors apart from 17 Two ACPI 15 Errors always come at the same time as the whole WHEA 17 stack These devices throw an error for instance, with multiple entries. Its 41 Errors in total. PCI\VEN_105B&DEV_E11D&SUBSYS_E11D105B&REV_00 PCI\VEN_10DE&DEV_22E9&SUBSYS_000010DE&REV_A1 PCI\VEN_144D&DEV_A80C&SUBSYS_A801144D&REV_00 PCI\VEN_144D&DEV_A810&SUBSYS_A801144D&REV_00 PCI\VEN_8086&DEV_272B&SUBSYS_40F08086&REV_1A

Basically everything on the PCIE Bus - some Intel stuff, then Nvidia, Samsung NVMe drives...Ill try to get a complete list.

This exactly. Its not just the card, its like the whole PCIE Bus at once. Like I have said, it has no observable effect on the usage of the machine, but the error logs are there. The machine can run 24 hours with no errors, but once you restart it or sleep / wake it, these pop up.

@MyPC8MyBrain Dont want to draw conclusions so early, but it seems like the problem is solved, even with the Advanced Optimus enabled there was no issue after updating the drivers & disabling PSR. What still bugs me is this (cant help myself). Flood of WHEA corrected errors. Happens during power state transitions, power source change, or on a clean boot. Has no impact on performance, no BSODs, but they are there. Just about every device that is on the PCI Express bus throws an error (the individual errors have different Primary Device Names). Should I just ignore them?

Sure thing, but it offers no dedicated solution, and the repairability (after the warranty period) is non existent due to CPU+GPU being both not only soldered, but also fixed by an epoxy adhesive (Lenovo does it like this for the last 10 years). BGA repairs in this case are, in most cases, impossible. Thank you for such an elaborate reply to this issue. I have performed a clean uninstallation of all graphics drivers via DDU, then installed the newest iGPU driver (from Intel, not from Dell), then the Quadro/RTX PRO driver. I have installed Intel driver only, without CC, it does not cause dGPU switching. Disabled PSR via Registry. Battery seems to behave normal, will observe the behaviour during the upcoming days.

Thank you for the suggestions. I knew that it has something to do with the iGPU driver, that the touchpad entry is not really the culprit. However, I dont know why is there no entry about iGPU / Display driver failure. These systems have Dell iGPU drivers that install without the Command Center. If you install the CC by yourself (through MS Store for instance) it results in a bugged optimus behaviour, where the machine keeps powering the dGPU on and off. The last iGPU driver that Dell provides is 32.0.101.6987 from 29.7.2025. There is no dynamic refresh rate on this machine (sorry, its in Slovak language, but you see the setting): Nvidia RTX Pro driver updated to the latest version as of this morning (581.80), the regular way with a clean install, not using DDU or anything similar. Intel GFX, I am going to try it, but as I have said, it causes other problems. Dell BIOS and EC package was updated about a week ago to the latest version, which is still the last available. If the issue persists, I will have to run dGPU only, yeah. In the Nvidia Control panel, the options are Automatic, Optimus and dGPU only. What exactly is difference between the Automatic and Optimus setting? AFAIK the Optimus always goes through iGPU, even when running dGPU, and the Automatic is like an on-the-fly MUX Switch. Or am I wrong? At the end of the day, the question still stands. Why has the issue manifested only after a month? I am using the machine in the same way every single day, and there was no recent alterations to the drivers or the configuration. The only thing that I have done differently, is that I have played Witcher III while docked, for like 10 minutes, to test temperatures.

The laptop has been working fine for a month, since this day, I have experienced this weird issue. Without a clear link to any activity or app, while browsing or watching a video, on battery or plugged in, the screen freezes, goes black (without backlit), then flashes about two or three times and stays black with backlit on. Like when MUX switching. However, after this, the laptop either freezes or continues to work, but without image. Have to force reset. Switching off Advanced Optimus in the BIOS seems to do the trick, but I would like to keep my battery life. Also, there was no issue for a month. The drivers are the same. I have experienced this several times during the last few hours. Each time it happens, there is this entry in the eventlog, exactly like this one: Anybody has experienced this behaviour ? Thank you.

No, I do not, and you still have to disprove that fairy tale of yours. Exuse me, but do you actually understand what the term ambient means? Because unless you are in a tropical climate, where room temperature is 31-32°C, then no, you are not idling 2°C over ambient. And even this was reached, by your own words: "(on Cool Power Plan in bios, on power saver in windows, if i switch to ultimate performance in windows the score will be over 24k, but i idle 10c hotter)" So thats that. But yes, lets all compare our meaningless idle values whilst we cripple our system in every way we can. Or it just might be that you do not know what idle is. And you still have to show the actually meaningful value - the core temperatures during the CB23 rendering phase. Not the 1 minute HWInfo window with power draw topped at 26W. This power draw, to be precise: The second thing you, apparently, have no idea about, is efficiency. Under no circumstances will the i9-12950HX be more efficient than Ultra9 285HX. It has nothing to do with its temperature. Efficiency is measured by comparing the value of input to the value of output, applied here it means electrical power in => compute power out. And the old Alder Lake CPU will not hold a candle to the Arrow Lake CPU, whatever PL you cap it at. They have almost the same PL1/PL2 values, yet the Ultra scores more than 50%, even when you hit Ultimate Performance. Thats efficiency. Brought in by the advancements of lithography and architecture.

Excuse me, but where have you been for the last decade? Laptop CPUs frequently go all the way to the TjMax since Coffee Lake came. Boost to the max => Throttle => Keep sustained levels. Its not pretty, but it is like this for almost a decade. Phone SoC chips behave the same way, consoles do too, as do Macs. Precisions behaved the same way, too (with TjMax at 100°C). On the PMP16/18, this spike is there for about a second, then it drops down to 80-90s. Keep in mind that the heat transfer from these delicate lithographies is difficult, even with a vapor chamber. Overall perfomance? That overall performance, which is currently the best amongst all mobile workstations ever designed, almost on par with gaming laptops that have way more power budget, and at the same time better than both its direct competitors? Do not mistake heat for temperature, its not the same thing. Efficiency has nothing to do with the operating temperature either. The cooling system in this machine is far superior to anything that Dell has ever produced. It is able to cool down the machine whilst keeping the fans more or less relaxed. There is no issue. All the review units that I was able to source the data from behave the same, even the NotebookCheck´s PMP16 and PMP18. I have no itentions of repasting the unit at the moment, if someone here will beat my benchmark results significantly after a repaste, that would change my mind. Idling at 90+° out of the box is not normal, but lets stay out of the fairy tales please. No modern Intel CPU, not even desktop, will idle at 2-4°C above ambient, unless the ambient is 30°C (if you live in tropical climate then you may be right). I am typing this from a workstation with 285K, cooled by a quad radiator system, and even this idles at 34°C (coolant at 27.8°C, ambient 21°C). In laptop it is simply impossible, unless it has been in a cold for very long and you have applied liquid metal. Pro Max Plus 18 idles at 38-40°C , and I am 100% sure that none of the old Raptor Lake Precisions will idle at lower values with the stock TIM. P.S.: It might be about time to stop judging thermal behaviour of a highly advanced semiconductor device against the boiling point of water, just because it was the standard back in the day. Yes, modern CPUs run hot, hotter than we would like, but thats the way it is. To cool a 150W CPU to a "nice" temperatures you need a pretty beefy heatsink even in a desktop.

I dont think so, these CPUs behave like this in laptops, it corresponds to the Tjmax. I belive that @AL123 observed the same behaviour.

I believe that there is an online review comparing PMP18 to PMP16 and the HP ZBook G1i Fury 18, the PMP16 does not lose a lot compared to the 18" variant, its within single % digits. The HP was noticeably weaker in most tasks.

Hello guys, I am enclosing the comparison between the Dell Pro Max Plus 18 and the ThinkPad P16 Gen3 with the same configuration. You may as well ignore the PL/TDP data, as on the PMP18 it means theoretical maximums of the platform, whereas on the TP P16 G3 it is couting on the platform power / thermal limitations already. Both systems were tested with clean OS and stock power settings, same conditions 1:1, it was just a misunderstanding. TL;DR: With a 55% larger platform power budget, the Dell has about 25% more power during the sustained CPU load, during the combined load of both CPU and GPU (such as gaming) the Dell has about 32% more power. It has also significantly lower noise output, but the CPU temperature does rise up to the 105°C and can spike up to 109°C on individual cores. On the ThinkPad, the temperatures peak at around 98°C due to the more aggresive thermal policy. Measured stock memory latency in this configuration at 108ns (PMP18) vs 151ns (P16G3)

To be honest, my work is not GPU heavy either, and the dGPU only sees actual usage during gaming, so I see where you are coming from. And yes, having the GeForce option would be nice, as a cost saving option for those who have no real usage for RTX Pro, like you or me. From my own experience and observations though (>20 years computer hardware, including laptop board repair), the Quadro / RTX Pro cards are just more reliable, by both the GPU itself and the memory. I dont know if they are better binned, or is it the fact that they tend to run lower voltages, more strict power limits and lower clocks, but they just work. I am not saying that they are immortal, but going back 10 years, I cannot really remember the last time I had a dead or problematic Quadro dGPU in my workshop. So I would pick the RTX Pro just for this reason alone, like I did. But of course, the GeForce equipped laptops outsell them by a mile, and they tend to be abused more, so that skews the data and I am aware of that.