The SpeedZone - Towers Performance: Power Mac G4/800 To Dual G4/1.42

Below you will find performance results for the currently shipping Power Mac Towers, and a some from previous generations. This will hopefully give you an impression of relative performances. We will be adding new machines on an on-going basis ... so check back once in a while.

The current crop of Power Macs were tested when running OS 10.2.3 .. the operating system they shipped with. Likewise the earlier machines were tested running the version of OS X they shipped with. However the scores for the earlier machines are relative, rather than absolute. The relative scores should be fairly accurate. Just keep in mind that the marks for the older machines may be 'in the ballpark', rather than spot on.

For a full listing of the differing specifications of each machines check the Power Mac Specs Pages

As you'll see in the performance results below, in some cases the 1.25 GHz machine from the previous series of Towers out-classes the current 1.25 GHz machine. This is especially true in terms of drive performance and when playing games. We think that this is due the fact that the old 1.25 GHz machine (which was the top of the line machine when it was released), has twice the L3 cache of the current 1.25 GHz Tower. It is also possible that the drives on the new Towers just aren't as fast as the ones that were in the last generation.

It is interesting to note that the previous 1.25 GHz machine cost $1,300 more when it was released than the current 1.25 GHz machine costs now.

If you are in a production type environment and have a strategy for utilizing all the power that this crop of Dual Processor Towers put at your disposal, the machines will pay for themselves. You should take note though that in most of the tests, huge amounts of processing potential was left unused by the applications we ran. For example our QuickTime DV encode only used 55% of the processing capability of the 1.42 GHz Mac. 45% of its processing power sat there twiddling its thumbs waiting to dance. To utilized these machines to the max you need to keep them fed with data. Our testing indicates this means running multiple jobs at once. This is now something that is quite easy and efficient to do with OS X. But it requires some planning on your part.

The tests below are from our suite of real world application tests. These tests feature a diverse selection of applications commonly used by the Mac community. The test suite was designed to render an accurate and well rounded picture of a machine's performance. All of the tests below (with the exception of the game test) were timed with a stopwatch. The times were then converted to percentages relative to the Dual Processor G4/1 GHz released in January of 2002. This machine is set to 100%. For all scores higher numbers are better

Desktop Tests

Booting Classic under OS 10.2 (Jaguar) is much faster than under previous versions of OS X. Booting Classic on the new single GHz machine took only 14 seconds, the 1.25 GHz machine about 18.

Mac OS X's new File Search function relies on both processor and drive performance. (Only 5% of the processing power of the 1.42 machine was used)

We have no explanation for why the last crop of Towers is beating the pants of the current ones in this test. Perhaps the L3 Cache runs at a faster speed on the last generation? If you have any ideas share them

In the single file test the the processor & caching capability of the machine is less of a factor. This result would seem to indicate that the drives in the new Towers (except for the high-end one) are just slower

The test above creates and destroys 1,000 windows. See the Let1kWindowsBloom site for more info. The two top-end Towers have the same graphics card

Processing power and drive performance are factors

The time Photoshop takes to open is gauged

Large Document & Database Type Tests

A Macro (series of complex actions) was run in Microsoft's Word program. Word is part of Office X. Only 50% of the processing power of the 1.42 Tower was utilized by the application

A Macro (series of complex actions) was run in Microsoft's Excel spreadsheet program. Excel is part of Office X Only 50% of the processing power of the 1.42 Tower was utilized by the application

This test takes place in a large AppleWorks document. The old 1.25 GHz machine comes out on top, we believe, because it has 2 MB of L3 cache per processor, whereas the current Tower, of the same clock speed, only has 1 MB. The 1.42 GHz Tower also has 2 MB of L3 cache.

Number Crunching & Rendering Tests

The Fractal program has been highly tuned to take advantage of the G4 and is precisely the type of work that the G4 was made for. It will also gobble up whatever processing capability is present. This is a good test for assessing the fundamental processing potential of each machine. Really highlights the processing advantage of a dual processor machines, and the G4 processor, when applications are tuned to take advantage of these features

Certain rendering actions in iMovie take good advantage of dual processors

Only 50% of the top-end Tower's processing capability was utilized

Only 60% of the top-end Tower's processing capability was utilized

BareFeats Photoshop Test Suite - mimics a production type workflow in Photoshop. In the single processor Tower, the processor was saturated with data. The 1.25 GHz machine had 25% left unused. Only 60% of the processing power of the 1.42 machine was used

Certain of the functions of Photoshop can take special performance advantage of the G4 processor. A dual processor machine runs rings around a the single processor one, in this test. (Only 75% of the processing power of the 1.42 machine was used)

The test above consists of Photoshop functions that do not take any particular advantage of the G4. The reason that the dual processor machines don't show more of a performance gap is that 50% to 55% of the processing power those machines was not being utilized. So having an extra processor was not much of an advantage ... unless you wanted to do something else at the same time.

Encoding/Decoding Tests

CD-ROM drive speed and CPU processing power are the factors influencing the scores above. Our assumption is that the CD drive speed is the bottleneck in the top of the line Power Macs above. For the top machine only 60% of the processing potential of the machine was being utilized. Meaning that it was not getting data fast enough to keep the two processors busy.

As you can see by the consistency of the numbers, this is straight processing work. Dual processors are not a factor as Stuffit only makes use of one

Multitasking

MP3 Encode, AppleWorks search & replace and folder copy are all carried out at the same time. If you need to have your computer carry out multiple activities at the same time, a dual processor machine will definitely get you there faster. The old 1.25 GHz Tower has faster drive performance and double the L3 cache of its current counterpart.

QuickTime Encode and iMovie Encode are both carried out at the same time. This is heavy-duty processing work

The MP3 Encode, Photoshop Normal and Word Macro tests are all run at the same time

Gaming

Giants is tuned to take advantage of the processor. Indeed, when run on a dual processor machine, you will see Giants fly ahead of the same game running on a single processor machine. It also helps to have a better graphics card. The new 1.42 GHz machine turned in 57.5 frames per second

4x4 EVO2 is a demanding racing game. Both of the top machines here have the same graphics card, with the same amount of video memory.

Past Performance Results for previous Towers