Slashdot Mirror
Using Safari Slows Your System?
sandoz writes "Macenstein has up an interesting article with some evidence that running Safari seems to slow down unrelated programs. While the speed with which a browser renders a Web page is an important measure, the difference between browsers is usually a matter of a few seconds at most. To my mind, a more important measure of speed is how a browser affects the overall speed of your system." Some responses to the article suggest that memory handling in WebKit may be the culprit. The Safari developers have already responded to this article on the webkit.org blog. They explain why the slowdown might be occurring and how it's (probably) already been fixed in the nightly build. And they request more minimal test cases.
Hey wow, a piece of software isn't perfect, and the developers are trying to fix it. This is an exciting new paradigm for programming. Thanks for keeping me updated!
One time I threw a brick at a duck.
I have a 5-1/2 year old iBook. Running anything slows my system down... : p
This guy's the limit!
Don't you realise Windows has this technology already - it's been slowing down unrelated programs for years! (Sorry, I know it's cheap, but I couldn't resist!)
biopowered.co.uk - catalytically cracking triglycerides for home automotive use since 2008. Just say no to big oil!
A few months ago, I switched to Firefox because I was convinced Safari was slowing down my system. Just this morning, I fired up Safari again - and it is at least three times as fast as Firefox. Don't know what I was thinking...
This space intentionally left blank.
I have both the nightly and the original Safari version installed. The latter leaks ram like crazy which tends to slow things down. You would think they would have fixed this ages ago. But they haven't. Try closing Safari periodically.
Another observation I have is that 1GB of ram is really only marginally adequate on my 2.16Ghz Macbook pro. If you have safari open, iPhoto open, and god forbid, a rosetta app (e.g. Word) open - you're waiting five seconds for windows to come up as disk gets paged out. Unacceptable.
Just wondering what other's experiences have been running the nightly code. I've been doing it with Firefox (and when it was Firebird for 2 years). But I've not tried with WebKit. Is it fairly stable, better rending of pages and faster?
There are a few sites that are noticeably slower on Safari. Its one of the only reasons I'm using Firefox. That and there are a few plug-ins that are better than Saaft and company.
I love safari and gladly use it over any other browser. However, since 10.4.5+, I have noticed that (as a whole) there appears to be an inability of OS X to free RAM up as efficiently as it used to. Programs like Safari, after many hours of usage, will remain as a HUGE RAM/virtual memory sink. I constantly quit Safari to try and alleviate/fix this.
But what seems to happen is that the process "kernal task" keeps eating up more and more ram even after Safari is shut down. After a couple days of usage, I feel the need for a restart just to flush out this annoyance.
Sure, in the grand scheme of things, It's only a minor annoyance, but it is definitely noticeable and something I hope is dealt with when 10.5 comes out.
my last sig was too controversial... now, a new and improved useless sig!
If a train leaves Boston with an I-MAC running Safari doing 40MPH heading west and a 2nd train leaves Chicago doing 60MPH heading East with an mini MAC running Safari and Quake 4. How many I-PODS will be sold funding more "Hi I'm a MAC" comercials before: a. My Karma can be kicked down another notch b. The code will be fixed ? c. This post will be changed to flame bait d. This post is silly and you can't believe your still reading it?
The S60 webkit is a port of Webkit to the Series 60 3rd edition platform. Nokia has created a memory manager for this port that can make the webkit works with low memory. If only I can have the low memory footprint browser in my Mac.
-- tinyhack.com
in the macenstein article they too noted that cpuintensive tasks like quicktime were not slowed but memory intensive tasks like photoshop were. Also they noted that the in memory and virtual memory footprints were several fold higher for safari than for firefox.
clearly this is a no brainier. Safari is using more memory and doing so in a demanding way. I don't know why but I assume it probably has something to do with how it handles the back-forward cache, fast page compoaition, and images. Maybe there's some memory leak too, since safari's offtprint grows during the day.
But this is utterly unsurprising. If you run a big memory app like photshop you already know better than to be running other apps that consume memory.
The only problem I've had with safari is not this but there are just some webpages that don't seem to comlicated that make it grind to a halt and use 60% of the cpu. One example is pricegrabber.com.
Some drink at the fountain of knowledge. Others just gargle.
Not only that, but the number of plugins available for Firefox make it really worth it. Adblock and Greasemonkey and Web Developer and Firebug give me functionality that's simply not available with Safari or WebKit.
I'm in the hole of the broadband donut.
"Since I upgraded to 10.4.8, Safari crashes on me about once a week. Forum advice was to run "repair permissions", I did but it didn't help. "
That seems to be the advice given for everything by some people. I'm not sure why anyone should think it would help in this case or many others.
I suppose you could try reinstalling the application after getting it off the install disc with Pacifist:
http://www.charlessoft.com/
before doing that you might also try removing Safari's preference file: com.apple.Safari.plist from the "Preferences" directory in your Home library, so that a new one is generated, in case there's a corruption in that causing problems. All this, again, is unspecific advice, but at least it's not Voodoo.
You might also like to try the Camino browser. That also uses Mozilla's Gecko engine, but has far better integration with the platform. It also uses a Safari-style bookmark manager:
http://www.caminobrowser.org/
All that Firefox lacks is password storage in the keychain and bookmark sync across multiple machines.
Between the two, it doesn't matter where I am, Safari has the same stuff in it.
Which is handy 'cause my stupid G4 iBook has a thermal fault somewhere around the NVRAM and I really should call the repair centre and see how they're getting along.
It's not like the G3 iBook had 6 bad main logic boards put in it....
(And yet I still can't stand to use Windows as a primary OS. Maybe I'll get one of those laptop Solaris machines. Or one of those old PowerPC Thinkpads that run AIX.)
I've yet to find a solution for the keychain password storage, but there's a plugin from Google called Google Browser Sync that I use to keep Firefox on my Powerbook and the mobile Firefox on my flash drive synchronized.
I'm in the hole of the broadband donut.
So when I read this item, I told myself "oh, so... what I'm experiencing isn't normal. it can be news in slashdot... wow." Firefox has different effects on different people I guess...
* Using a clean profile + a nightly build doesn't help. Submitted bug reports do not get any interest from devels except tagging it with "perf" (I know, they're busy, but look - it's news on slashdot when it's Safari on Mac).
bugs in question? so far, I was lazy enough to file just these: 366728; 368365; 368908; 369044; 369682; 370697
pls don't reply w/ "worksforme". I spent considerable time trying to not reproduce the slow down effects, as you might guess...
From the article:
<quote>
The only thing different was that I had been surfing the web a bit while the render was going on that day, where the day before I had not. "Surely surfing the web on a mulit-processor machine shouldn't add 15 minutes to a render", I thought. Well, yes it does actually, if you're using Safari.
</quote>
<p>Put another way: "Surely letting the computer ONLY do my render won't be any faster than letting the computer render AND surf the web". Surely you jest? Doing a standalone render vs. doing (anything else) while rendering should make your render take longer. If it doesn't, file a bug with Adobe and ask them why they aren't utilizing the hardware properly.</p>
It didn't help because the people who gave you that advice are imbeciles who believe in voodoo. It was able to help with occasional problems caused by Classic and bad installers circa OS X 10.1, but there are hardly any circumstances where it will make any difference to anything on more recent versions, as explained here.
Using HTML in email is like putting sound effects on your phone calls. Just say <strong>no</strong>.
Opera for Windows and Mac has been free as in beer for quite some time now. Go download it from Opera.com, it's pretty good.
Sunwalker Dezco for Warchief in 2016
Safari slows down other programs so that it can look faster in comparison. So that's the secret of their success!
I'm not an apple fanboy yet, but I'm really impressed with the immediate response of the Safari development team. Imagine if IE was slowing down some other program--the last group you'd expect to hear from would be the IE dev team--so far outside the realm of possibility as to be laughable.
This story isn't about some exciting new program, or some cool technology, or even an updated version of a popular piece of software. This is a report about a non-critical bug for a web browser. A web browser with limited market share and a number of good alternatives. Oh, and the developer has already acknowledged it, explained it, and described some of the progress being made to fix it. This isn't about technology, it's about a minor decision in the production of a web browser not being the optimal solution.
Slashdot covers a huge range of topics, a lot of information goes through it each day. It's constantly bring in new editors, and they get craploads of submissions each day. Keeping all of that in mind, I just fail to see how anything in this article was worthy of a front page spot on the site. I'm not calling for the firing of the editors or anything, just making fun of them a little bit for posting something silly.
One time I threw a brick at a duck.
I used to use Saft for ad blocking and saving sessions, but Saft slowed down Safari even more. Now I use SafariBlock for blocking ads, and Forget Me Not for saving sessions, and it works fine.
You can import adblock filter lists into SafariBlock, as well as add site on the fly.
I'm gonna do what I want and I'm gonna get paid -- Tom Waits
What is not always realized is that there are two Safari versions.
Safari 1.x runs on OS X 10.3 and earlier.
Safari 2.x runs requires OS X 10.4.
All the performance fixes go into Safari 2.x only. The only fixes for Safari 1.x are for security problems.
I'm running OS X 10.3, and I have noticed that Safari has severe severe performance problems. I think it is caused by Flash applets, which bring it to its knees. Websites with flash applets frequently will cause 100% cpu usage. And trying to use flash based video players often can't keep up, because it is demanding more than 100% cpu.
So now that there are so many flash ads, exploring the web is a painful experience.
* * *
OK, I know this is futile, but I'll try to head off the obligatory responses:
Q. Do you really expect Apple to enhance Safari for older operating systems?
A. I don't expect them to. But I wish they would fix the performance problems.
Q. If Flash ads are the problem, why don't you block them?
A. The ads pay for the web sites.
Q. Why don't you upgrade to OS X 10.4?
A. Because that would break certain other software needed for work.
Q. Why don't you use another browser?
A. I do.
Q. Why don't you just get a get a new, faster computer?
A. Because I keep expecting Apple to release a new model, in between a Mac mini and a Mac Pro.