Separate names with a comma.
Discussion in 'General Mac Discussion' started by CubaTBird, Sep 18, 2004.
the score is 105?!?? my ibook gets 100.2 !! this is seriously wrong, im at an apple store right now
LOL, man, you need to put in another stick of 512MB and change the energy management to maximum from auto.
Either that, or Xbench is a notoriously unreliable piece of rubbish.
I'm going to go with that explanation.
However, it does seem that setting processor performace to highest has a very noticable effect on performance, outside of just increasing the xbench score.
you know whats sad though, the proc performance should have been automatically set to highest, its kinda weird that apple wants their desktop comps running slower than they actually can... it would make sense for a laptop to conserve energy but this isn't a laptop...
Alot of things can change the score on Xbench. You have to set CPU to the highest, close all apps besides Xbench(I close windows as well to be safe), and make sure nothing will interfere with xbench. I got a 99.93 on a rev b 12" pbook 1 Ghz. But, I killed the HD test and got 113.13. Notice on how much the HD removes from your score. I tryed every test one without of each. The HD effected it the most.
I've been saying that for months now, but people are still using it as a benchmark for Mac performance. Is there no other program that can accurately measure Mac performance?
Just for the fun of it
Check out my "real world" bench test elsewhere in this section. It's entitled "Dual 1.25 too fast".
I used iDVD background encoding performance as a testing tool, as this is the app I use most and would want to see the most performance boost.
Just for your knowledge I am currently thinking about writing a benchmarking app in cocoa, which will be all optimized for 64 bit and altivec testing and all that good crap. I am hoping it will turn out to be better than xbench .
I will start a forum about it if i do decide to write it.
I wish there was. I've been thinking that perhaps an AppleScript of some sort which will load up a suite of apps, perform a set of operations, and then record the total time taken would be nice. However, that would require that the system being tested have all the same version of all the same apps installed, so in practice it would probably be impossible to implement.
Maybe as a community we need to talk seriously about constructing some standardised application benchmarks (using AppleScript maybe to drive the whole shebang) that people who own the relevant apps can download and execute...something like the PSBench series of actions for benchmarking PhotoShop performance, and set up a centralised repository for submitting and searching the benchmark results. Sites like Barefeats are good, but I find it a little frustrating sometimes that they don't always run the same tests on every machine, and their methodolgies are sometimes a little...err...'uncontrolled'. xlr8yourmac.com has some searchable benchmarks too, but I think the community over there is quite a bit smaller than the one here, probably mostly due to the fact that registrations for their forums have been closed for years.
I dunno...what do people think? Am I just talking about re-inventing the wheel here? Is there already a good site which allows downloading of standardised application benchmarking scripts and then has a good searchable database? I've always thought that actual application benchmarks (eg: benchmarks like UT2004 Flybys, PS7Bench, iTunes album encode times, FCP clip render times) with standardised run-scripts and test data, were better than purely synthetic benchmarks.
Opinions? Flames? Donations to my PayPal account?
Maybe there should be a MacRumours benchmark, or NusuniAdmin benchmark written by NusuniAdmin. I'd happily use that instead of XBench. Every time someone uses XBench scores as irrefutable proof that one system is clearly faster than another, I just laugh.
Actually I was thinking of a app called XiBench or something similer....dunno about you but that sounds nifty to me. Anyways I wouldn't be writing it for a long time anyways, i am currently under heavy development of an other app.
how do you change the processor performance?
maybe they don't set it to maximum because of heat issues in that case? maybe if the processor is at the highest setting, the cooling system can't keep up and the processor will eventually overheat..
The Processor Performance setting is under the Energy Saver preference pane in System Preferences. It is under the "Options" tab.
They ship the computers with the energy saving features turned on in order to save energy. In many parts of the world, that is a far more important consideration than xbench scores.
It's only on laptops right? cause FuzzyBallz's post sounds like hes saying more ram should be added to the imac and increase the processor speed.
maybe i'm just confused and read his post wrong
The feature was originally introduced with the Titanium PowerBook G4 DVI models (the 667MHz and 800MHz models), but is now in all modern Macs, including desktops. It extends battery life in laptops, and reduces energy use and heat generation in desktops. The G5 PowerMacs and iMacs both have this feature, and you can set the processor to highest, automatic, or reduced performace.
I take it my G4 iMac shouldn't have it then... which ends my confusion.
I can understand why they have it in the desktop, lots of people us it just for typing etc and it saves power, but why not ship with it set to best performance?
Tut Tut apple *slaps wrist*
This is also a demo mac... as in one on show in an apple store... therefore... it's been played with endlessly.. messed about... only 256mb ram... probably hasn't been restarted today... loads of stuff running in Virtual memory, etc... If you benchmarked that machine after a restart... or benchmarked it out of the box... it'll be around 120 not much more though.
i restarted it and had a guy who worked there type in the admin password... it also had 512, which should be enough i mean seriously.. if my ibook can keep up with an imac g5, seriously something is wrong there... i went and benched a dual g5 and got a score of around 187.. which made more sense
I think we had this discussion back when the G5 first came out. Nothing was different from automatic and highest.
not totally true. I remember readin this from a developer doc or something in xcode that talks about the settings, all automatic does it if the machine gets too hot or whatever it will slow it down, or if not that much power is needed it will slow down. WHen you run benchmarks and such it will not affect the actual benchmark (at least it shouldn't).
You know, it strikes me that this forum should be about the leaps made with the G5. Unfortunately it is about how to measure performance and what affects performance. If that's what we're talking about... my guess is that the speed difference just isn't there. Then again, we're talking about miniscule differences. I mean for us e-mail, internet, movie watching, itunes listening, occasional video slapping together, proficient base users, what makes a machine feel faster is the UI speed. For me to be ecstatic about the speed of my computer and OS, here is my wish list:
I click a dock item and it appears, instantly.
I switch between windows in a browser and they appear, instantly.
I navigate through files and folders, they appear instantly.
I pop a DVD in and DVD Player begins, instantly (the wait seems to be an eternity).
I tell my computer "Phone for Jim Appleby" and it understands me, every time.
I go to an internet page I've visited before and it appears, instantly.
And it makes brewed coffee, instantly.
No disrespect to your ibook, but it most certainly cannot keep up with a 1.8 G5. I think everyone needs to throw Xbench out the window. Just wait until www.barefeats.com performs some real world tests. They will show it is about 95% of what a single Powermac G5 1.8 is. (my guess anyway)