Hacker News Reader

Comment by hendersoon | Hacker News Reader

This smells fishy to me.

Why did they test in High Performance power mode? The default is Balanced.

High Performance mode doesn't allow the CPU to downclock, keeping it at the maximum frequency at all times, even with zero load. It also maintains a higher screen brightness, and stops wifi and disks from entering power-saving modes too. This is NOT a negligible difference.

The _concept_ behind optimizing for energy savings in a browser is absolutely brilliant. Why pollute it by using such blatantly fake numbers?

View on HN

Using HPM eliminates confounding factors. You won't get real world numbers because some of the benefit is provided by the OS/hardware, but you'll know if your changes are actually making a difference during the times that they are active.

It's simply not an appropriate test. You don't use high performance mode when you care about battery life.

I don't get your objection here. Not only does high-performance mode eliminate OS-level confounders that would otherwise render consumption deltas highly questionable, it does so in a way that produces worst-case results.

What exactly are you complaining about?

All that other stuff impacts battery life in a huge way. Those variables matter. You can't just remove them, call it a "worst case scenario", and walk away. That is not valid.

When a reputable site like Anandtech reviews a laptop, they don't force it to maximum performance mode, because nobody that cares about battery life would ever do that. It's testing a scenario that _doesn't matter_.

In fact, not only is it not the default, but maximum performance mode doesn't even _show_ by default-- you need to expand a "show additional plans" dropdown in the power control panel to see it. Opera went to the trouble to do that, and it was not by accident.

I think you're looking for a followup study that complements these results, but that study would be not as valuable as with these to inform their interpretation.

They state in the article that there are a lot of variables to account for when measuring battery savings - keeping it in high-performance mode likely makes it a more reliable control for testing against.

Again, testing battery life savings with all power-saving options disabled is simply not appropriate. It stinks.

Reply on news.ycombinator.com