4 ms·
One factor that is important to me when comparing browsers is resource consumption. I don't like it when my fans spin. I wonder if the "time" tool that comes w
by mg 5y ago
One factor that is important to me when comparing browsers is resource consumption. I don't like it when my fans spin.
I wonder if the "time" tool that comes with Linux is a good way to quantify it? When I do "time firefox reddit.com", wait until the page is fully rendered (including the ads) and close the browser, I get:
time firefox reddit.com
real 0m13,089s
user 0m9,411s
sys 0m2,882s
Does that mean that Firefox used about 12s of CPU time to render the frontpage of Reddit? (I guess user+sys is the amount that counts)
"time firefox news.ycombinator.com" gives me about 7 seconds.
- nsonha 5y agoyou are counting bootstraping/closing the browser. Check firefox devtools timeline for a detailed breakdown.
- mg 5y agoYes, that is what I want to look at. The real resource impact of starting the browser, looking at a website and closing the browser. Using an in-browser tool to measure that would be like asking a robber how much they stole.
- cmeacham98 5y agoMost people don't close and open their browser between page visits, so time spend opening/closing the browser is a magnitude of importance less than time spent loading the page. I'd rather have a browser that takes 10secs to start and 0.1secs to render a page than 2sec to start and 1sec for each page render. Additionally, timing externally like this ends up indirectly measuring unrelated metrics like the speed of your DNS server and internet download speed. I've seen people with particularly shoddy ISP DNS experience a 0.5-1sec swing in page load times after the DNS entry gets cached.
- mg 5y agoI think DNS speed will not impact the "user" and "sys" values of time. Only the "real" value. I think user+sys give the amount of CPU time used, not impacted by wait time.