Hacker News new | ask | show | jobs
by npongratz 4181 days ago
> When away from a power socket, I use Safari instead of Chrome. Saves battery a LOT.

Were you comparing Safari with 0 extensions vs. Chrome with 0 extensions? Or does each have a different extension load?

I will often fall back to Safari if a page doesn't render properly in Chrome, but that's because I installed a bunch of extensions on Chrome and have consciously left Safari pure as the driven snow. It's a different problem domain, but I know certain extensions can use resources in surprising ways.
1 comments
schrodinger 4181 days ago
One factor is that I believe chrome has flash by default, and Safari does not
link
STRML 4181 days ago
It's not just that. I see nearly 2x better battery life using Safari over Chrome on a 15" rMBP, and I have plugins turned off (click to play). It just uses a lot less CPU, full stop. I don't know what the real difference is, but I suspect Safari just does a better job of idling than Chrome does.
link
nkurz 4181 days ago
My semi-verified guess is that a lot of the difference is CPU usage when two-finger scrolling on large pages. I just tried it on this page with OSX 10.9 and a 15" Mac Pro, monitoring with Activity Monitor and Coconut Battery while bouncing rapidly up and down.

Chrome burns 90% of a CPU on one thread, and another 40% on another. Coconut reports about 32 Watts used. Safari maxes at about 12% plus 8%, and tops at about 15 Watts. Base usage on this machine with the current screen brightness and applications that happen to be open is 10-12 Watts. I get approximately the same results with Incognito mode, which suggests the problem is core Chrome rather than add-ons.

link
rmckayfleming 4181 days ago
When you click the battery dropdown on the top right, Chrome is consistently listed as an "Apps Using Significant Energy". On previous releases of OS X I also noticed that by default it would use my dedicated GPU at all times, whereas Safari would not (not sure if this is still true).
link
STRML 4177 days ago
This is no longer true, but it still does request the dGPU on certain pages, such as the Chrome Web Store.

What's especially obnoxious about that behavior is that if you disable the dGPU, it works just fine. So what's the point?

Also, one would expect Chrome's own Web Store to work well in Chrome... but Google's attitude toward long-standing bugs is anything but exemplary.

link