Forums | MacLife

In an interesting quirk, we found some dramatic speed differences in these systems depending on how we set the new Processor Performance preference, found in the Energy Saver pane of System Preferences. By default, Processor Performance is set to Automatic. Changing the setting to Highest improved our results quite a bit. Our MPEG-2 compression test, which took 9:44 on the 1.8GHz system at the stock setting, took 8:32 at the Highest setting. Overall Speedmark scores on the 1.8GHz system rose from 214 to 225 when we switched from Auto to Best.

That's the interesting part....

This should quiet things down in here a little:

Much better Photoshop 7 comparisons.

Note that the best times are in red and also note which machine has the red numbers most of the time.

Later, Frank

I need to point are few things out here.

1. Keep in mind that, as someone else said, these MA results were by their own admission, last minute tests and not comprehensive.

2. I think that MA Photoshop testing has ALWAYS been completely wacky. I can use almost the same machine and try to duplicate the test and my times are always better. Frequently dramatically better (2x and 3x!)

I totally disregard MA photoshop benchmarking. I don't know what the trouble is, maybe they are in some kind of force field generated by the Maximum PC folks down the hall.

As I mentioned earlier, I have a dual 2gig on order (it will be a few weeks  ) that will be configured with 1.5 gigs ram. As soon as I get it I will run these tests and post the results.

Also, I disagree about the donkey nuts comment (as funny as it is). My G4 933 has regularly bested various PCs of a comparable clock speed and higher (2.4ghz P4). But I test with an action that runs a series of filters and processes in Photoshop so it tests all the factors. Personally I think the MA method of run one test, relaunch (or reboot, even) is ridiculous. Nobody uses any app that way.

This should quiet things down in here a little:

Much better Photoshop 7 comparisons.

Note that the best times are in red and also note which machine has the red numbers most of the time.

Later, Frank

Much better test results *lowers gun from head*

The demo machines in Apple stores should be fat with RAM. There's no excuse to keep those demos at 256/512 - they are there to show what Macs CAN do, right?

They most likely have the machines configured that way so that people can test them in their stock configuration. Now, it has been a long time since I was in an Apple store (closest one is 4 hours away) but I think I remember that while most of the machines were stock configurations their were a couple of machines that did have extra RAM, etc.

Later, Frank

Those PS7Bench numbers are much more believable.
...Dual 2.66GHz Xeon.

That's far more impressive, and truthful.

I agree. That is my expectation (in terms of performance.)

The questions remains: what is MA doing that causes numbers so totally out of line with everyone else?

And the bigger question is, why hasn't anyone responded since Rik did last weekend and why do they still have that stupid article/test results posted on the web site?

Later, Frank

This should quiet things down in here a little:

Much better Photoshop 7 comparisons.

Note that the best times are in red and also note which machine has the red numbers most of the time.

Later, Frank

Much better test results *lowers gun from head*

The demo machines in Apple stores should be fat with RAM. There's no excuse to keep those demos at 256/512 - they are there to show what Macs CAN do, right?

They most likely have the machines configured that way so that people can test them in their stock configuration. Now, it has been a long time since I was in an Apple store (closest one is 4 hours away) but I think I remember that while most of the machines were stock configurations their were a couple of machines that did have extra RAM, etc.

Later, Frank

I can see wanting to test the stock setup on most every other Mac, but if you're going to buy a tower you probably have in mind adding more RAM. The bigger issue is - shouldn't Apple have upped the base RAM to 512MB in the entry tower and 1GB in the Dual? Well, they should have at least one machine at the store that's MAXED with RAM - I'd love to see what the Dual G5 does with 8GB of RAM!

OK everyone. To lay this to rest you must check out the thread at Arstechnica with results from PSBench. It has solid results from the G5's and compares them to pretty much everything from a Pentium II 266 all the way up.

Cutting to the chase: The G5 dual 2.0 blows everything right out of the water. (The Dual G4 1.4 hangs in there pretty well, too, I might add)

I'm getting my stock G5 tommorrow, a week exactly after MacAddict. I'll be able to run some tests then using Adobe Photoshop 7.0.1 and Mac OS X Jaguar 10.2.7, as well as Panther 10.3 (7b53).

-It better come tommorrow, or my pencil will get it...in half.

Well IT DID COME TODAY !

whoooo HOO!

Well it has these specs, so they will be a ball-park figure when compared to the other G5 towers.

Specs:

1.6Ghz G5
ATi Radeon 9600/64 8x
256 MB DDR
Combo Drive
80 GB HD

I have plans put two 512's in and an ext. 120GB + 80 GB Hard drive setup. I just didn't wanna spend any more money than I had to, so I figured upgrading later would be a good idea.

This machine is awesome, inside there are no visible wires (except slightly in one corner.) The motherboard is blue and the machine is bigger than I thought it would be. It's cool because you can take off the side panel and just leave the clear plastic wall up, and you can see all of your components with out worrying about whether they are getting hot or not. It's like having a "window" like on the Win/Lin box's that are out there. I put some LED ANTEC lights inside and it doesn't even need to be modded to look badass as some of the other modded cases out their. Only you have the option of covering up the window at any time.

When you take the window off, the fans speed up and the machine is almost as loud as a small vaccuum cleaner, oh and red warning light shoots on to warn that you better put that window on or you will get a swift kick in the arse. Put it back on and the machine isn't even audible in a normal room. I had people come and look at it and they were surprised that it was even on.

Hey, if anyone wants to see Photoshop scores on it, I'll post them later tonight. I'm at work now and can't get to it yet...I should have brought it here to work. I'll also post pictures of the small mods I'm doing to it on my website.

[edit:] maybe rik is posting mods? in their mag??

Ok, I "give".

It has become apparent from times posted here and in the "Circuit Board" thread that with 512MB ram or less, high-end G4's and G5's suffer very badly when running Photoshop with files that are about 100MB or so in size, or larger.

I imagine it was the intent of Macaddict's chart to demonstrate that fact.

Ok, I "give".

It has become apparent from times posted here and in the "Circuit Board" thread that with 512MB ram or less, high-end G4's and G5's suffer very badly when running Photoshop with files that are about 100MB or so in size, or larger.

I imagine it was the intent of Macaddict's chart to demonstrate that fact.

See, this is one of the reasons I was somewhat annoyed having tried the 1.8GHz/512 setup at the Apple store. Unless you can afford upwards of 4GB of RAM (or more), you'll eventually hit the hard drive on large, multi-layered files. I expected the new Serial ATA, higher system bus, etc. to speed up disk-intensive processes significantly. This just didn't happen - it just felt like my old G4 system once the hard drive had to kick in. Seems there's still no substitute for a SCSI scratch disk.

Perhaps there are still some technical issues with Serial ATA hardware that Apple needs to work out (?). I still feel my PC does better when Photoshop runs out of RAM and must use the hard drive. Maybe it's just my imagination. 

What Macaddict probably did was run Photoshop set to use 50% ram in all the tested machines. I asked Rik if this was the case earlier in this thread but got no response from him. This gave Photoshop only just over 200MB ram to play with. That being said, the scores are still awful. The number of history levels may also have been set to 20.  I ran the Photoshop trial with 6 history levels.