Welcome to the MacNN Forums.

breadiu · Apr 13, 2005, 03:45 PM

From the benchmarks posted by MacTouch, it looks as if Tiger is significantly slower than Panther.

In the past, each major iteration of OS X has shown significant speed improvements. It's rather disappointing to see such lackluster performance out of Tiger's initial release.

azdude · Apr 13, 2005, 03:48 PM

You must be reading the data wrong, then.

Longer bars are better. Compare two identical machines... the Tiger equipped machine is at least equal to, but almost always scoring higher for all machines and all tests.

Yippee! How the hell does Apple do it?

xi_hyperon · Apr 13, 2005, 03:54 PM

Yeah, it was misread, as the benchmarks are looking very impressive with Tiger.

Quite nice.

juanvaldes · Apr 13, 2005, 04:06 PM

So mainly improvements in the UI and I guess some refining of the memory code for the G5s. Not the same kind of grand sweeping improvements like before but still not bad and we knew it would not last forever.

Socially Awkward Solo · Apr 13, 2005, 04:19 PM

Just as I though, all those Tiger is 100x faster turns out to be hype. On a G5 the speed improvements are next to none except on UI.

Eug Wanker · Apr 13, 2005, 04:20 PM

So is that with Tiger 10.4's Quartz 2D Extreme enabled?

Originally posted by Socially Awkward Solo:
Just as I though, all those Tiger is 100x faster turns out to be hype. On a G5 the speed improvements are next to none except on UI.

I think OS X UI speed improvements are VERY important. Looking good!

Socially Awkward Solo · Apr 13, 2005, 04:24 PM

Originally posted by Eug Wanker:

I think OS X UI speed improvements are VERY important. Looking good!

I spend less time worrying about resizing windows manually than I care about threading and switching between apps.

On my dual G5 with 1.5 gigs of RAM if I am in photoshop with a 100 meg image for an hour and switch to safari I heard the hard drive grind for 20 seconds BEFORE safari even comes forward.

Eug Wanker · Apr 13, 2005, 04:26 PM

Originally posted by Socially Awkward Solo:
I spend less time worrying about resizing windows manually than I care about threading and switching between apps.

On my dual G5 with 1.5 gigs of RAM if I am in photoshop with a 100 meg image for an hour and switch to safari I heard the hard drive grind for 20 seconds BEFORE safari even comes forward.

You need more RAM. 1.5 GBs is way too little for what you do. BTW, Photoshop CS 2 breaks the 2 GB barrier too.

Socially Awkward Solo · Apr 13, 2005, 04:38 PM

Originally posted by Eug Wanker:
You need more RAM. 1.5 GBs is way too little for what you do. BTW, Photoshop CS 2 breaks the 2 GB barrier too.

Ya but I mean a 100 meg image isn't that bad. you are supposed to have 4x the RAM for the image you are opening. I think 1.5 gigs is enough for a 100 meg image.

But down the road I will get some more.

Millennium · Apr 13, 2005, 04:50 PM

Originally posted by Socially Awkward Solo:
Just as I though, all those Tiger is 100x faster turns out to be hype. On a G5 the speed improvements are next to none except on UI.

Isn't that where it's needed? The UI has been the main performance bottleneck in OSX for a long time. Speed it up, and everything else will speed up accordingly.

I'm a bit concerned about the Quartz benchmark, though. According to this, the PMG5 1.8 took a nosedive performance-wise when going to 10.4, but the iMacG5 1.8 got faster, to the point where it got the fastest results of all the configurations in the test. It seems awfully strange to me that one machine would get faster and the other not.

CatOne · Apr 13, 2005, 04:58 PM

Originally posted by Socially Awkward Solo:
Just as I though, all those Tiger is 100x faster turns out to be hype. On a G5 the speed improvements are next to none except on UI.

100x faster? Where was that?

50% faster in UI is nothing to sneeze at. And none of the tests run were on a dual CPU box. Wonder why not? Seems a pretty crappy representation of hardware, but in general it's faster in nearly EVERY case... I think that's pretty damned good, if it varies from 1% to 23% to 50%.

hmurchison2001 · Apr 13, 2005, 05:13 PM

I expect Tiger to get a bit of a boost with subsequent updates to 10.4.1 and 10.4.2.

Also there's not much of a way to test the fine grain locking features until you get your hands on it with a DP system. It's obvious that Tiger hasn't gotten slower despite the new plumbing. I'm not complaining.

Socially Awkward Solo · Apr 13, 2005, 05:13 PM

Originally posted by CatOne:
100x faster? Where was that?

Sorry I meant to say a Zillion times.

JHromadka · Apr 13, 2005, 05:28 PM

Looking good!

zerostar · Apr 13, 2005, 05:43 PM

Don't forget dashboard is running and the new spotlight services and all the new demons, etc. are all running the background by default.

Being faster or even on par as Panther with all the extra [tiger] stuff running is a good thing� IMO.

CatOne · Apr 13, 2005, 05:43 PM

Originally posted by Socially Awkward Solo:
Sorry I meant to say a Zillion times.

Thinine · Apr 13, 2005, 05:51 PM

XBench scores mean jack ****. All I can say is that, in my usage, Tiger has been noticeably faster than Panther in every way. And I never noticed it being slower.

Big Mac · Apr 13, 2005, 08:59 PM

Originally posted by Socially Awkward Solo:
Ya but I mean a 100 meg image isn't that bad. you are supposed to have 4x the RAM for the image you are opening. I think 1.5 gigs is enough for a 100 meg image.

But down the road I will get some more.

Yeah, as others have pointed out, your example has nothing to do with multitasking performance. Safari was swapped out to disk. And I have never heard extravagant claims about Tiger's increased performance.

OreoCookie · Apr 14, 2005, 05:57 AM

Tiger isn't out yet, how are we supposed to benchmark it?