[ab-in-va]

I benchmarked an iMac G5 at the Apple Store and it scored 113! The iMac was a 17" display model with 512MB RAM, all the applications closed. I find this score to be kinda low. I did not run Cocktail or anything like that on it before benchmarking it, so I presume this score is how it would do out of the box.

Anyone got any tips on improving benchmark scores? What programs do you use to increase system performance?

[the_glassman]

The score seem very low, in most cases lower then the eMac. Somebody has already stated in another thread that xbench isn't the most accurate, or say all be all of bench marking. Did you try Cinebench?

[PKRADD]

Bench scores do vary from machine to machine and depend on what App you're using and whether it is properly optimized for a G5 chip. I understand no current benchmark Apps are at the moment. Several bench scores have ben posted elsewhere and they vary greatly. In the end it means nothing. What matters is how the computer runs when you use real world programs.

[Luca Rescigno]

XBench currently has three submitted benchmarks from the new iMacs:

107.50
112.35
149.62

Lots of variance there, but it seems as though the CPU test and the disk test are the ones holding it back. Funny to see the mighty G5 processor falling down so bad when they stick it in a dinky iMac with a slower bus speed. And all of them are using Seagate hard drives... nothing wrong with Seagate, and the model listed (ST380013AS) should be just fine, with a 7200 RPM speed and 8 MB of cache. Perhaps the disk test is so slow because it's being mounted vertically, not horizontally.

[ab-in-va]

Originally posted by Luca Rescigno:
XBench currently has three submitted benchmarks from the new iMacs:

107.50
112.35
149.62

Lots of variance there, but it seems as though the CPU test and the disk test are the ones holding it back. Funny to see the mighty G5 processor falling down so bad when they stick it in a dinky iMac with a slower bus speed. And all of them are using Seagate hard drives... nothing wrong with Seagate, and the model listed (ST380013AS) should be just fine, with a 7200 RPM speed and 8 MB of cache. Perhaps the disk test is so slow because it's being mounted vertically, not horizontally.

cool! i stirred up a real sh*t storm on my first post

[toti]

Originally posted by Luca Rescigno:
Perhaps the disk test is so slow because it's being mounted vertically, not horizontally.

I have seen this mentioned a few times in the iMac discussions. Where on earth do people get these crazy ideas from ? Can somebody sahre a link from a hard disk manufacturer stating that vertically mounting a hard drive will result in lower perfomance, please !

I have gone through ALL reading material Seagate has to offer on the disk in the G5 17" iMac, and there is NO mention whatsoever that vertically mounting the drive might make a dent in its performance.

What is much more likely is that the benchmarker left the machine in a powersaving state ( spin down harddrive whenever possible ), had the "automatic" speed selection of the CPU active ( which also affects the rest of the machine in a non-generous way ) or had something running that in some way or other disrupts the tests. ( Of course these are GUESSES and I do clearly state that neither do I have an iMac to verify these results, nor do I have enough knowledge on the circumstances when the tests where made - BUT I DO SAY THESE VERTICAL MOUNTING SPECULATIONS ARE UTTER UNTIL PROVEN OTHERWISE )

Gawd how ignorance can really pinch the marrow

[Finrock]

I decided to take a trip to the Apple Store at International Plaza. They didn't have any 20" iMacs in stock but they did have two on display.

I got a chance to use it for about 15 minutes. After the first few minutes I was rather nonplussed. With 512 MB of ram (an identical matched pair - according to Apple System Profiler) I was hoping for a little more performance. I went to the System Preferences and saw that the Processor Performance (Energy Saver - Options) was set to Automatic. I asked the friendly Apple Store Employee if she would be so kind as to change that to Highest.

Boy did that make a difference!

I was truly impressed with how fast and smooth everything felt, from GarageBand to iDVD, Safari to Word. I was amazed.

I bet a ton of the xBench scores posted around the net haven't taken this into consideration. I would have run an xBench test before and after but there were too many people and I was already hogging the machine.

[Luca Rescigno]

Ah, I see.

BTW, I said "perhaps" because I really didn't know. I am not even sure if vertical mounting actually has any effect on things, I was just speculating. But there's SOME reason for the poor disk test, and you're right, it's probably due to the energy saver settings. By default, they spin down the hard disk while it's not in use, which is a bad thing. Constantly spinning down and spinning up puts more strain on the disk than just leaving it spinning all the time. And it does greatly decrease performance because if you try using the machine after a few minutes of inactivity, it has to spin up, which takes a few seconds.

I'm thinking that last test (the one that got about 150) is much more in line with how the iMac G5 should perform. It's too bad that Apple's default settings (not to mention the default RAM loadout) make for such poor performance. I'm afraid there will be a lot of disappointed iMac G5 users in the next few weeks.

[toti]

Well, we'll just hope that most of them read some forums where they can get sound advice on what to do then, eh ? At least they'll see their new machine 'spring to life'

I am still waiting for my machine. I hope it arrives before october ( they make no promises at the local Apple centre tho.. ) and once it arrives, I'll be sure to run some benchmarks of my own

[AndrewP]

Originally posted by Luca Rescigno:
Perhaps the disk test is so slow because it's being mounted vertically, not horizontally.

It makes no difference as long as the platters are either parallel or perpendicular to the surface of the Earth. Anything off axis will stress the bearings resulting in pre-mature failure, but not necessarily any performance issues.

Note that most (if not all) "enterprise-class" disk arrays and servers mount disks vertically (for more efficient cooling).

Also, I've noted vastly different Xbench disk results if I run multiple tests in a row. I'm assuming that this has to do with buffer and cache hits vs reading directly from the platter.

[CWeider]

LOL Finrock, I drove up today from Tierra Verde to International Plaza as well. Wasn't as crowded as I was anticipating, but I did have to wait to get some keyboard time on the new iMacs. It's refreshing to have a weekend without evacuation orders & hurricane watches.

Regards,

Chris

[naphtali]

Originally posted by AndrewP:
It makes no difference as long as the platters are either parallel or perpendicular to the surface of the Earth. Anything off axis will stress the bearings resulting in pre-mature failure, but not necessarily any performance issues.

Note that most (if not all) "enterprise-class" disk arrays and servers mount disks vertically (for more efficient cooling).

Off-topic, but does this mean that using your PowerBook with things like the Podium CoolPad can cause problems?

[macaddict0001]

Originally posted by naphtali:
Off-topic, but does this mean that using your PowerBook with things like the Podium CoolPad can cause problems?

yes but its not serious enough to consider doing anything about it. for example the fan is all the way vertical and how many people do you know with a failing fan.

[MrForgetable]

my 15 inch PB has a higher score than that o_O

[Lancer409]

wow ... in the store .. it was pretty snappy ...

what's a pb 12 rev c, with 1.25gb of ram bench at?

[Lateralus]

XBench is worthless guys.

[turtle777]

Originally posted by PowerMacMan:
XBench is worthless guys.

So what's worthwile ?

-t

[RonnieoftheRose]

It's a myth that vertically mounted drives slow down. The spindle turns so fast that the effect of gravity is virtually nil. Jobs invented this little story to defend the second generation iMac's optical drive whacking the screen and keyboard. Nothing more. It certainly doesn't apply to hard drives.

If the XBench score is low it is because the G5 needs a fast bus to shine.

[Shaddim]

You did turn the processor speed to "highest" in System Preferences/Energy Saver didn't you?

[turtle777]

Originally posted by MacNStein:
You did turn the processor speed to "highest" in System Preferences/Energy Saver didn't you?

We should have a sticky for that...

-t

[Mrjinglesusa]

Originally posted by Finrock:
I decided to take a trip to the Apple Store at International Plaza. They didn't have any 20" iMacs in stock but they did have two on display.

I got a chance to use it for about 15 minutes. After the first few minutes I was rather nonplussed. With 512 MB of ram (an identical matched pair - according to Apple System Profiler) I was hoping for a little more performance. I went to the System Preferences and saw that the Processor Performance (Energy Saver - Options) was set to Automatic. I asked the friendly Apple Store Employee if she would be so kind as to change that to Highest.

Boy did that make a difference!

I was truly impressed with how fast and smooth everything felt, from GarageBand to iDVD, Safari to Word. I was amazed.

I bet a ton of the xBench scores posted around the net haven't taken this into consideration. I would have run an xBench test before and after but there were too many people and I was already hogging the machine.

This makes a HUGE difference! This was on a 933 MHz 14" iBook:

81.2 w/ custom Energy Savings
91.2 w/ Highest Performance

That's why it is so hard to compare XBench scores - who knows what kind of settings people have.

[chrisutley]

FWIW, I haven't had a chance to put any RAM in my 20" yet, but with 256MB in there it just scored 154 on XBench. Can't wait to load her up with some memory and have some fun.