Chrome is still a memory hog on macOS compared to Safari
Long been derided as a resource-consuming web browser, a recent test by a developer reveals Google Chrome to use multiple times the memory of Apple's Safari in macOS.
Chrome is often the source of ire for users who find the Google-created browser to be bloated and too keen to consume a Mac's available memory. In many cases, users are pointed in the direction of the lightweight Safari, but in a new test report, a developer shows how bad Chrome is on RAM.
A blog post by Flotato developer Morten Just spotted by iMore attempted to find out how much the memory disparity is between Safari and Chrome. Two tests were performed, with the browsers experiencing minimal load alongside a more realistic scenario.
For the minimal load test, a virtual machine was set up with a clean macOS installation, and the browsers were made to open up two tabs showing Twitter and Gmail. A snapshot of the memory and CPU utilization was captured 250 times per second using psrecord.
In the minimal scenario, the average RAM usage for Chrome for just the Twitter tab was 730 megabytes, ten times the 73 megabytes observed for Safari. Flotato, an app for loading mobile versions of websites as an app on macOS, shaves the RAM usage down a bit further to 63MB.
Under the two-tab test, Chrome's memory utilization reached 1 gigabyte after a minute, while Safari stayed consistently below 80 megabytes.
![The memory and CPU consumption of Chrome (left) and Safari (right) when opening 54 tabs in macOS [via Flotato]](https://photos5.appleinsider.com/gallery/40422-77863-flotato-memory-consumption-54-tabs-chrome-safari-l.jpg)
The memory and CPU consumption of Chrome (left) and Safari (right) when opening 54 tabs in macOS [via Flotato]
Instead of a virtual machine, the second "stress test" used Just's own macOS installation and opened up 54 tabs in each browser. While Safari kept the average amount of RAM used per tab to a svelte 12 megabytes, Chrome's average per-tab memory usage was a massive 290 megabytes.
Both Apple and Google regularly improve their respective browsers, with Safari 14 introducing more performance improvements alongside new features like a tab bar redesign, customizable start page, and privacy-related elements.
To Google's credit, it has been working to introduce some enhancements to change how it handles background webpages, improving processing performance as well as reducing the load on a Mac's battery, and even a version built for Apple Silicon. Memory consumption has been a long-running problem, which Google has repeatedly tried to address, but evidently it is still an issue.
Chrome is often the source of ire for users who find the Google-created browser to be bloated and too keen to consume a Mac's available memory. In many cases, users are pointed in the direction of the lightweight Safari, but in a new test report, a developer shows how bad Chrome is on RAM.
A blog post by Flotato developer Morten Just spotted by iMore attempted to find out how much the memory disparity is between Safari and Chrome. Two tests were performed, with the browsers experiencing minimal load alongside a more realistic scenario.
For the minimal load test, a virtual machine was set up with a clean macOS installation, and the browsers were made to open up two tabs showing Twitter and Gmail. A snapshot of the memory and CPU utilization was captured 250 times per second using psrecord.
In the minimal scenario, the average RAM usage for Chrome for just the Twitter tab was 730 megabytes, ten times the 73 megabytes observed for Safari. Flotato, an app for loading mobile versions of websites as an app on macOS, shaves the RAM usage down a bit further to 63MB.
Under the two-tab test, Chrome's memory utilization reached 1 gigabyte after a minute, while Safari stayed consistently below 80 megabytes.
The memory and CPU consumption of Chrome (left) and Safari (right) when opening 54 tabs in macOS [via Flotato]
Instead of a virtual machine, the second "stress test" used Just's own macOS installation and opened up 54 tabs in each browser. While Safari kept the average amount of RAM used per tab to a svelte 12 megabytes, Chrome's average per-tab memory usage was a massive 290 megabytes.
Both Apple and Google regularly improve their respective browsers, with Safari 14 introducing more performance improvements alongside new features like a tab bar redesign, customizable start page, and privacy-related elements.
To Google's credit, it has been working to introduce some enhancements to change how it handles background webpages, improving processing performance as well as reducing the load on a Mac's battery, and even a version built for Apple Silicon. Memory consumption has been a long-running problem, which Google has repeatedly tried to address, but evidently it is still an issue.
Comments
I was using Chrome for quite a while, but I noticed I was hearing the iMac's fan a lot. I installed the program MenuBar Stats which gives all sorts of system info- CPU and RAM usage, reading on the almost 30 temperature sensors inside the machine, and of course, fan speed.
Well, the fan would often go to 100%. The CPU temp would soar to 165-170º.
Then I switched to Safari. Not only did everything load faster, but the fan stopped coming on. The programs that were operating at the same time, worked faster. The CPU temp rarely goes about 145º.
Chrome simply sucks.
Does the M1 version of Chrome have the same problem? The article doesn't say.
No, this is not poor programming, it’s a purposeful design choice that Google uses to improve resilience and fault tolerance. Google also has a more open plug-in/extension model. Isolating plug-ins/extensions into their own process and memory address space within a tab keeps a poorly behaving, e.g., memory leaker, plug-in from bringing down the entire Chrome application if something goes wrong. Google made a choice and they chose resiliency and fault tolerance over memory efficiency. This makes perfect sense for their design intentions and as users it gives you a choice if you have to use or access certain sites or use certain plug-ins that are not as stable as you’d like. There are times when I'd rather have a single tab crash than have the entire browser crash. This may be more of a concern on Windows, but it's not a defect, just a characteristic of the application.
I've had to make these kinds of design choices and there have definitely been valid reasons for choosing out-of-process over in-process. Other reasons for running out-of-process include separating the lifetime of your server process from the lifetime of your clients, maintaining state between client sessions, and keeping your service/app running and doing things in the background. Yes, it costs more in memory consumption but most virtual memory based systems including Windows can handle low memory situations reasonably well, i.e., allowing the system to slow down rather than running out of memory and crashing. It's up to the designer to choose the model that best fits their design intentions and you'll find a mix of in-proc and out-of-proc services on all OSs that support these models. Google took one approach and Apple took another. No big deal in a world where you can have as many browsers on your computer as you would ever want to have.
If you “have” to use Chrome for a particular site it should not present a problem unless doing so requires you to open multiple tabs and you have so little memory that you get into a heavy paging situation. It’s not like you get a rebate or any material benefits from having a lot of free RAM just sitting there doing nothing.
I, personally, set Safari and Chrome to open the same 5 tabs at start, and if you count ALL the Safari processes, it's actually MORE than Chrome.
This article is just untrue.