[deathandtaxes]

I just got my new iMac G5 20" on Thursday. It's fantastic; REALLY quiet. Bigger and heavier than I expected, but the tilt and base are excellent. I wanted to check the performance using Xbench for comparison to my older iMac FP 17". The numbers seem low for a 1.8Ghz G5; consistently 116. In fact, I got the same 116 score with the stock 256megs of memory AND after adding a 1gig stick. On the Xbench web site, some folks are posting scores as high as 160 with EXACTLY the same configuration.There seem to be two groups of scores. Many in the 112-116 range, then they jump to 148-160. What's the difference. I don't get it.

[Londor]

System Preferences > Energy Saver > Options > Set Processor Performace to Highest.

[el_humpo]

Dual channel memory might also have an effect. If you can, try installing two identical sticks of RAM and see what happens.

[Finrock]

Its the energy saver option, I guarantee it. See my original post here:

http://forums.macnn.com/showthread.p...48#post2193424

-Finrock

[PEHowland]

Originally posted by el_humpo:
Dual channel memory might also have an effect. If you can, try installing two identical sticks of RAM and see what happens.

Don't know about on Apples, but on PC's dual-channel memory is 99% hype and 1% substance. You are generally lucky to see more than about 5-10% improvement on memory benchmarks in dual-channel mode and even less on real world applications. It's obviously worth enabling if you're going to buy two sticks of memory anyway, but it's not actually a particularly significant performance driver. See this article and part two of the article for a discussion of this in the context of PC hardware.

Maybe things are different on the G5 chipset.

[Pierre B.]

How many times has to be told that Xbench is not appropriate to test the performance of a system? There should be perhaps a sticky entitled "Xbench is junk", or something like that, so that people take note before start feeling dissapointed about the performance of their machines.

[RogerR]

Out of curiosity, is that Energy Saver-Options-Set Processor Performance setting peculiar to G5 Macs? I ask because I notice my G4 iMac has no such setting.

[Finrock]

G5 Macs, PowerBooks, iBooks... these all have the above settings for Processor Performance in OS X.

[deathandtaxes]

Thanks. That made a huge difference. I bumped the performance option under Energy Saver to highest (instead of automatic) and the score jumped from 116 to 157. Wow!! I wonder if keeping it on highest will have any noticeable effect or cause the fan to cycle on more often.

[RogerR]

Originally posted by Finrock:
G5 Macs, PowerBooks, iBooks... these all have the above settings for Processor Performance in OS X.

But non-laptop G4s don't have the setting, correct?

[yikes600]

I'm confused. Why would Apple would have adjustable processor performance in a desktop?

[deathandtaxes]

Correct. It's on the Energy Saver/Options screen on the iMac G5 but not there on the iMac G4/FP running the same version of OSX.

[deathandtaxes]

Perhaps someone more knowledgeable about the G5 can comment but I would assume the G5 is smart enough to cycle faster or slower depending on demand, which would allow less heat buildup. I guess the G4 isn't as smart.

[RogerR]

Originally posted by deathandtaxes:
Correct. It's on the Energy Saver/Options screen on the iMac G5 but not there on the iMac G4/FP running the same version of OSX.

Thanks.

[Lancer409]

... felt snappy in store ..

[turtle777]

Originally posted by yikes600:
I'm confused. Why would Apple would have adjustable processor performance in a desktop?

Maybe we'll get a snap-on batterypack to use the iMac as a laptop...

-t

[cambro]

You should leave your performance setting on "Automatic" unless you are doing processor intensive stuff all the time. The dynamic adjustment is good enough not to make a bit of difference in most real-world applications.

Contrary to what you might think, a computer left on consumes plenty of power. Apple allows you to save a little power with virtually no noticeable performance hit. Although we're in a country that buys SUV's, HUMMERS, and consumer appliances that suck power even when off just because nobody gives a *hit about conservation, conserving a little power when CPU cycles aren't necessary is good for lots of reasons.

[Matt OS X]

Originally posted by deathandtaxes:
I just got my new iMac G5 20" on Thursday. It's fantastic; REALLY quiet. Bigger and heavier than I expected, but the tilt and base are excellent. I wanted to check the performance using Xbench for comparison to my older iMac FP 17". The numbers seem low for a 1.8Ghz G5; consistently 116. In fact, I got the same 116 score with the stock 256megs of memory AND after adding a 1gig stick. On the Xbench web site, some folks are posting scores as high as 160 with EXACTLY the same configuration.There seem to be two groups of scores. Many in the 112-116 range, then they jump to 148-160. What's the difference. I don't get it.

I xbenched my imac g5( 20", 256mb, 160GB) and I think I got 151.95 as results. What does this supposed to mean? good? fair? I'm new with xbench stuffs, Thanks!

[ratlater]

Originally posted by Matt OS X:
I xbenched my imac g5( 20", 256mb, 160GB) and I think I got 151.95 as results. What does this supposed to mean? good? fair? I'm new with xbench stuffs, Thanks!

Xbench is really pretty useless. The only thing it's really useful for is comparing the same system. So you could compare your score to other 20" iMac owners and see if more RAM or other changes would help your performance.

It's misleading to judge an iMacs score to another systems score though.

-matt