The Story You Want to Read, and the Actual Story

Posted by Matt Birchler
— 1 min read

This article has been making the rounds the past week or so:

I bet you have more than 2 tabs open right now, I sure have more. So here's a stress test where I open 54 tabs while measuring the impact on my Mac's RAM and CPU.

The article goes on to show Chrome using x more RAM than Safari. Hence, the Safari fan club starts posting this wildly.

Then some context:

This post is unfortunately nonsense, because it's making a fundamental measurement error: not including renderer processes for Safari, and only measuring the browser process (which should be expected to be roughly O(1) memory in the number of tabs).

Ah, so Safari’s numbers are so much lower because the test was not counting the other processes that Safari spun up to run those 54 tabs. This certainly explains why as more tabs were opened, RAM usage stayed basically level.

It sounds like this is an accident on the original author’s side, but still, this thing was all over the place recently, and as that same Hacker news commentor pointed out:


I have half a mind to try a similar test of my own to see the results with full counting of Safari’s other processes. I would expect Safari to still come out ahead, but we’d have to do the test to be sure.