Hacker News Reader

Comment by savant_penguin | Hacker News Reader

Idk I expect my computer to be able to run at 100% without some sort of death clock

Blaming bad thermal design on "use case" does not make sense to me.

If you'd like to run some intensive numerical programming on your laptop just because, I would find it outrageous for it to shutdown due to thermals. Can other hardware designers use that excuse too? "Oh your software use case is not covered by the cpu voltage regulators", "the power supply was not meant to be on all the time", "the power cable should not be connected for that long", "you should not use 90% of your ram for 4 hours straight, it was not designed for that use case"

"Oh the average end user only uses the car 1h/day. This 5h long trip was not the expected use case, that's why it overheated"

View on HN

Duty cycle- the ratio of time a component may be powered- is really quite common.

Cars can only rev the engine in the red zone for a limited duration. Lasers on CNC machines, welders, engine hoists, etc. all have specific capacities that you shouldn't exceed.

Expecting your CPU to run flat-out at 100% indefinitely without reprocussion seems strange to me. Heat is the great killer of electronics, and unless your rig is water cooled I don't know that I've used a laptop before with the thermal management to withstand extended full load- either it throttles down or it damages itself.

Throttling is fine in a laptop. I think the more concerning part is the oscillations. The CPU should reach a steady state where it can provide consistent performance at 100% load.

Is there a difference when averaged out over time?

peak heat?

With modern "burstable" CPUs, aren't they always peaking for short periods of time? Aren't there different heat thresholds, some that are more dangerous than others?

Like maybe you could change the throttling curve from (percent "max") 120-120-100-100-80-80-80-100-100... to a more consistent 90-90-90-90-90... but as long as the CPU knows internally when to throttle down to avoid damage, does it really matter either way? Fast-fast-slow-slow vs medium-med-med-med?

it depends on how hot it gets in each sequence, but it seems to me the m2 laptops get hotter than the m1 laptops, and if this OP is to be believed, reach a higher peak temperature that's over 100 degrees c.

> ...I don't know that I've used a laptop before with the thermal management to withstand extended full load- either it throttles down or it damages itself.

Setting your "it damages itself" remark aside for the moment, every laptop I've ever personally owned runs flat-out indefinitely at its rated maximum speed. I make a point of choosing laptops that have properly-designed cooling systems.

Electromigration is real, but that's a process that happens over many, many years. It is (_strictly speaking_) damage... but the only real way to entirely prevent it is to never use the device in question.

> ...unless your rig is water cooled...

In my personal experience, for mainstream (and, yes, this includes "workstation" and "server") CPUs, water cooling isn't notably better than a big-ass Noctua heatsink with a couple of their 140mm fans strapped to it.

Phase change "heat pipes" are really fuckin good for rapidly transferring heat from the contact pad of these heat sinks out to their radiating fins, and the big fans that are usually paired with them are quite good at cycling enough air through the system to ensure that the heat is moved out of the radiator in a timely manner.

> every laptop I've ever personally owned runs flat-out indefinitely at its rated maximum speed.

You must be using a very nuanced definition of "rated maximum speed", incorporating time-dependent effects like Intel Turbo Boost, and CPU power limits that vary depending on power source, OEM-specific software options, and possibly discrete GPU usage. Which makes your claim almost a tautology.

Other products ARE designed with use cases in mind. Take your car to a track and see how many hours you can push it for before it dies. Bonus points if it’s electric.

> If you'd like to run some intensive numerical programming on your laptop just because, I would find it outrageous for it to shutdown due to thermals. If you get a laptop without a fan and expect to push it like this then don’t be surprised if it can’t keep up.

Again, is the thing you want to do with the laptop within the scope of that the designers meant it to be used for.

That being said, I find it strange that their performance oscillates rather than hits a steady state. That part they should fix in a software update.

This specific laptop's thermal issues aside, a cpu that throttles under constant full load can also be seen as a cpu that gives you extra performance for a limited time. That kind of "boosting" is standard now even on desktop processors, but makes the most sense for laptops, where cooling capability is restricted by size and weight concerns.

Just wait until you find out about SSD write endurance

Yeah, a year and a half of near-constant writes, adding up to (if my math is right) 5000 -> 10,000 whole-drive-writes before failure.

<https://techreport.com/review/27909/the-ssd-endurance-experi...>

This is totally, legitimately incomparable to a CPU throttling to a fraction of its rated steady-state speed within a couple of minutes of constantly-applied workload because the laptop manufacturer decided to prioritize chassis thinness over cooling system effectiveness.

Reply on news.ycombinator.com