The power draw is being limited, but not because the CPU or GPU are running too hot (they’ll be in the mid 80s or even 70s). It’s when the power delivery parts (inside the laptop) get too hot to keep up and it can’t keep up. You can override those with programs like throttlestop, but the battery will drop MUCH quicker. When it’s hot the PL1 and PL2 drop to about 25 watts which is basically unusable on 11th gen i9, but it easily has the head room for 45-55 watts. GPU is largely unaffected which is weird. I’ve seen it get limited to around 60 watts, but the 3080 mobile below 80 watts is also awful.

For monitoring power usage I use hwinfo 64 in windows, I’m not sure if the portable version would work.

P1 gen 4 with the i9 and rtx 3080. Pay close attention to the power levels under heavy load. It will drop massively under long term heavy loads to try to prevent the battery from discharging. My machine only takes in a little over 170 watts from the power supply, but with a laptop cooling pad it can easily sustain over that 170 watt mark. It doesn’t happen instantly, it starts when the laptop is fully heat soaked (takes 30+ minutes with the cooling pad). You won’t notice it until about an hour or two in, but once it starts it will start accelerating as the battery heats up. Shorter loads that the laptop is more designed for it handles it just fine. It’s only when you push it for too long and too hard.

Also whats the power consumption of the mobile 4090 like sitting on but idle? Random programs trigger my 3080 for no reason and that GPU draws about 20 watts minimum. I want to upgrade, but I’d lose vram if I got anything less than the 4090 and I don’t know if I want all of that excess power draw when the system can barely benefit from it, and it makes using it as a laptop awful.

Well I hope you don’t like ThinkPads…

There’s nothing wrong with the device, Lenovo has confirmed this, and both motherboards my laptop has had have the same “problem”. This isn’t my only machine like this either, 16" Intel MacBook Pros are also known to discharge under full load, but that’s because they’re limited to 100 watt USB C.

There’s a reason why those devices run at minimum clock speeds when their battery is sub 5%.

It even happens with power efficient devices. All Macbooks will run at their lowest clock speed with a dead/low battery (even my M1), My Thinkpad T14 with an ULV CPU and it’s odd. It tries to limit total system power to around 25 watts, even though I have a 100 watt power supply connected. My theory is that since 30 watts is the lowest power supply it will run off of it’s trying to keep that 5 watt buffer. Unfortunately that means my CPU runs at 800mhz doing anything but idling. Laptops with dGPUs often just wont work at all, or are so far limited they’re unusable.

Some older laptops like my Thinkpad X220 will run at 800mhz on a 65 watt charger, but on a 90 watt charger it will run at full speed. But unfortunately in the days of USB C that makes things a lot more difficult.

Power consumption. Especially with turbo boost power consumption can easily spike well above what the power brick can deliver, so the battery is used like a capacitor. Or shit even without the spikes chargers can’t keep up. My laptop will actually discharge under full load with the full 240 watt charger.

It’s not normally an issue on REALLY low end devices (sub core i, like pentiums or atoms), but anything high end will reduce it’s power consumption without a batter installed.

If you remove the battery it will either A not work or B run extremely slowly. Always have a functional battery in your laptops.

Ideally find a way to limit the charge of the battery. But if you can’t nuking your battery is better than running at 800mhz or whatever your lowest clock speed is.

Yeah an unclean reboot shouldn’t break anything as long as it wasn’t doing anything when it went down. I’ve never had any issues when I have to crash a computer unless it was stuck doing an update.

They still consume way more power than the pi. My pi fully loaded uses less power than my N100 router idle.

I’ve been running one for the past 6+ months with no issues.

Only one device needs to be able to negotiate it.

Just about everything made in the last 20 years is capable of doing this.

If you don’t need a lot of GPU horsepower besides the AI stuff then you could just use the integrated graphics and have a dedicated GPU for the AI stuff.

Having multiple GPUs in your system isn’t really that special. Plug HDMI into GPU1 to make GPU1 drive your display/play games. Plug HDMI into GPU2 to have GPU2 do stuff. If you’re doing AI work then you don’t need anything connected to the GPU, the program just needs to know it’s there and to use it.

The only thing to look out for when using the iGPU and a dGPU is that the bios doesn’t turn off the iGPU if it detects the dGPU. If you have 2 dGPUs then it shouldn’t matter outside of maybe the bios wanting to use the first one.

It frees up sata slots for your massive array of hard drives.

Don’t just look at temperatures though, look at clock speed too. 95c+ is normal for modern high end CPUs (AMD 7000 series actively try to run at that temp under full load). What you want to make sure is that it’s not throttling.

If this is a server and you don’t want your thermal paste to be toast in a year then I’d suggest lowering the maximum temperature in the bios if it lets you.

And make sure it’s an actually good PSU too.

I know in gaming, possibly in other loads Nvidia 40 series, and especially 30 series love transient spikes which can easily exceed 2x the nominal power consumption. Make sure your PSU can handle those spikes both in terms of brevity, and current.

Try making sure you’re looking for a DSL modem, and not a DSL modem router combo. If it’s just a modem then there is no bridge mode to enable because it’s always in bridge mode. You’ll have to supply your own separate router, but that gives you the freedom to run whatever you want on your router while the modem does it’s modemy things.

If you already have a server and don’t need more resources then just leave it. No need to waste more power “just because”.

That said you could hang onto it until you have a need. You never know when a machine will shit the bed and having a half top is useful.

What makes you think that can’t happen to something just because it’s open source? And from all companies it’s from Canonical.

It’s “Selfhosted” not “SelfHostedOpenSourceFreeAsInFreedom/GNU”. Not everyone has drank the entire open source punch bowl.

why are you stuck with their questionable open-source and ass of a kernel

Because you don’t care about it being open source? Just working (and continuing to work) is a pretty big motivating factor to stay with what you have.

I’d pay more money for a non raid enclosure.