2 ms·
A lot of people here are criticizing the methodology. That's fair, I don't know enough to weigh in. But what I can say for sure is that in real-world use, on M
by Wowfunhappy 6y ago
A lot of people here are criticizing the methodology. That's fair, I don't know enough to weigh in.
But what I can say for sure is that in real-world use, on Mac, Safari absolutely feels lighter weight than other browsers. The fans spin up less, the battery lasts 30% longer, and other apps seem to be more responsive. I don't know if that's memory or cpu or what have you, but it's real.
You do pay a price in actual page-load times. Chrome absolutely has a small but noticeable edge in raw browsing speed.
I suspect this comes largely from Safari's tight integration with the rest of macOS. I recently had occasion to read a bit of Chromium's Cocoa UI code (to help the developer of https://github.com/blueboxd/chromium-legacy https://github.com/blueboxd/chromium-legacy track down some bugs), and my very surface-level interpretation of what I saw could be summed up as "how can we get rid of macOS's native functionality and use our cross-platform stuff instead?"
- eyelidlessness 6y agoI have plenty of anecdata to share on this confirming Safari is generally lighter, and for common usage that should be pretty reliable. But from the same anecdata experience, what happens when it’s not is awful. So the significant difference between Safari’s and Chrome’s model is how tabs are pooled into subprocesses. Safari quickly groups same-origin tabs into the same process which has a huge memory cost advantage. Chrome very aggressively spawns subprocesses per tab up to some pressure metric before it starts sharing processes across tabs, which consumes more memory but much more readily frees it too. Safari’s model is basically more shrewd and probably better for average users, but the pathological case is GBs of unused memory with unclear priority and total system stalls. Chrome’s pathological case is much easier to reach but much easier to recover from because for the most part closing or killing a bad tab process fully relieves its resource pressure.