5 ms·
Don't have a Mac so can't look into it but can anyone confirm this test is actually right? E.g. Brave claims (I don't use Brave I just know they had this claim)
by zamadatix 6y ago
Don't have a Mac so can't look into it but can anyone confirm this test is actually right? E.g. Brave claims (I don't use Brave I just know they had this claim) Safari is worse for memory usage https://brave.com/brave-one-dot-zero-performance-methodology-and-results/ https://brave.com/brave-one-dot-zero-performance-methodology... while this product that uses Safari claims it's 10x better.
Obviously both claims can't be true. From what I remember with Edge (pre Chromium) and Firefox I'm going to guess Safari doesn't really use 1/10th the memory of everyone else - maybe I'm just not familiar though!
- lucideer 6y agoAnecdata: I find browser RAM usage varies enormously depending on which specific websites you use. The Flotato test uses Twitter and Gmail (for the first set; it doesn't specify the sites used for the 2nd set). Both Twitter and Gmail are large, complex, slow webapps which likely do some sort of advanced active memory management (extensive use of service workers, local storage, etc.). They also load mainly (exclusively?) 1st party resources. The Brave tests use a large varied set of sites[0], which seem to be a combination of simple news & ecommerce sites. There are no complex webapps in the list. Modern news sites do have a reputation for being slow and heavy, but this will mostly be due to 3rd-party resources from ad companies; the sites themselves are simple and lightweight in terms of 1st-party content. So they seem to be comparing very different browser usage patterns. Given that, and looking again at the Brave tests: Safari does actually have lower memory usage than Chrome for most of their graphs. It just creeps slightly ahead of Chrome in the later tests. [0] https://github.com/brave-experiments/browser-comparison-toolkit/blob/master/scenarios/fullset.txt https://github.com/brave-experiments/browser-comparison-tool...
- ASalazarMX 6y agoAs a side note, GMail memory usage has improved vastly (probably because it was monstruous to begin with). Years ago my GMail tab consumed almost 500 Mb, but nowadays it uses less than 100, with my same usage patterns.
- mayli 6y ago300MB onload, dropped to 240MB after idling for few minutes. Are you sure it's using less than 100MB? in comparison with the post, it takes about 500MB ram, which is weird for such simple webpage. This HN page takes 40MB.
- ASalazarMX 6y agoOh sorry, I measured it on Firefox. Chrome on Linux started at 160 Mb, and it settled around 150. I've disabled Hangouts, and the right sidebar, so that might account for the differences. Funny how Firefox reports a third of that, although IDK if Chrome and Firefox task managers are directly comparable.
- chrismorgan 6y agoFour years ago, I found Gmail to tend to use 80–200MB in Firefox, mostly in the lower half of that range. Certainly I found the figures in this test utterly unbelievable for this reason; Safari ain’t lighter on memory than Firefox, let alone that much. But also for reference, it’s not like something of Gmail’s complexity has to use that much memory. Fastmail provides equivalent functionality minus chat in a 10–15MB budget, and you can add calendars and contacts (which are separate in Google land) and still keep it to 16–25MB.
- lucideer 6y agoGmail specifically has massively improved recently for me, but I feel it tends to be more of a spikey graph than a steady trendline over time. Memory usage went through the roof when they first rolled out their current redesign, despite already being far too high before. I assume Gmail will get worse again at some stage. Similarly, Google Docs & Sheets are abysmal currently; not really sure if they were ever significantly better or worse in the past. In terms of webapps, Google's approach has always seemed to be to make things that are powerful, sometimes usable, but with no focus on performance, cross-browser compat, or general industry best practice at all from a frontend engineering perspective. Any engineering quality in Google products has never in my experience lived in client-side code. This is exacerbated by the fact that "code quality" drives in other companies are often sneaked onto managerial priority lists under the guise of SEO requirements, something Google have never needed to meet.
- moksly 6y agoI use a MacBook Pro 2016 everyday. Safari is easily the best browser memory and systems wise, but I still mainly use chrome because it’s better at addblocking and synchronises with my non-Mac devices. Chrome is fine though, as long as you don’t open 20+ tabs. Beyond that it’s not fine. I replaced Firefox with Chrome because it kept ranging from a great to terrible user experience with various updates. This is all anecdotal and based on personal experience. On iOS I mostly use Safari even though that is kind of against my synchronised argument, but it works alright as I have found that I use my phone for very different web things.
- CharlesW 6y ago> …I still mainly use chrome because it’s better at addblocking and synchronises with my non-Mac devices. (In case you're interested…) iCloud for Windows now does bookmarks and passwords sync. https://support.apple.com/guide/icloud/icloud-for-windows-overview-mm6be2394337/icloud https://support.apple.com/guide/icloud/icloud-for-windows-ov... For ad blocking, I really like 1Blocker. NextDNS is great as well.
- moksly 6y agoI am, thank you.
- kitsunesoba 6y agoThe linked blog post doesn’t state it explicitly, but I suspect that the non-Brave browsers were not equipped with any kind of adblocking for the test. Brave has built in adblocking, so if my suspicions are true the test is rather disingenuous. At minimum Firefox and Chrome should have uBlock Origin installed and Safari should have Wipr or 1blocker installed to make it fair.
- strmpnk 6y agoI’d suppose some of it relates to sites with heavy ads. Safari with no extensions will likely require more resources which makes the comparison a little skewed vs Safari with some ad blocker support added via extensions. It looks like their test includes plenty of ad heavy sites like news articles which seems to confirm that the results are cherry picked.
- runako 6y agoI don't have stats handy, but this has been published lots of times by many different authors. As others have said, comparing browsers fairly is difficult. But Chrome is simply not an efficient browser on Macs. Anecdotally, it uses more memory and measurably consumes more energy. The Mac's fan will run more when using Chrome, and you will need to charge more often. If you use a lot of tabs, your entire computer will slow down. There's market validation for this assertion. There is a YC-backed company currently working on tech to offload Chrome tabs to the cloud so the browser uses fewer resources. This is not something that would be identifiable as a problem in Safari. If you have access to a Mac, you don't need to take anyone's word for it. Just use Chrome for a few days and then use Safari for a few days. The difference is pronounced. If you use your browser heavily, you will notice. Since password managers are in the news, Safari also syncs passwords to iOS devices (password sync also works in native apps). It's free.
- dahart 6y agoChrome does use a lot of memory, but I’m extremely skeptical of this blog post’s methodology and results. Does it seem realistic to have 54 different real web pages open that total under 1GB? (If they’re opening tiny fake pages, this test is misleading and useless.) Most real web sites transfer several megabytes on load. CNN.com loaded 4.2MB for me. Amazon.com loaded 4.3 MB. That’s just the data transfer of compressed assets. The images and content have to be uncompressed and rendered, which takes a lot more memory. I’m playing on a Mac right now and see nothing like what is reported here. When I open cnn.com in Safari, that tab alone takes 1GB. Chrome reports far less usage than that. Also “Real Memory” is not the same as active memory in RAM, it’s not necessarily giving a complete picture. Memory can allocated but unused or compressed or paged out.
- wooger 6y agoWe already know apple's own software has special exemptions in the OS e.g. from packet sniffers. Any reason to trust the ram usage for Safari nowadays?