The Performance Edge: Photoshop, Gaming And More - A Further Exploration Of The Comparative Performance Of The New Dual Gigahertz And 933 MHz G4 Towers

Wednesday, February 27, 2002

Lots of us are still using some version of OS 9. If you are a Photoshop user, for example, trying to run it through Classic when booted into OS X is like a bad Acid Trip. Most of the hardcore Games too are not quite yet ready for the move to OS X. So if gaming is your thing you will be spending a lot of time in the old, familiar Mac OS. OS X is getting more comfortable, and for some people essential, every day. But for a sizeable portion of Mac users it is still not quite there yet.

Below we take a performance look at the new Dual GHz and 933 MHz Towers (comparing them to each other and the last generation of Tower machines), when booted into a variant of Mac OS 9. The tests below confirm our perspective that, unless you are making heavy duty use of processor intensive applications (such as Final Cut Pro, Photoshop, iMovie, audio applications, etc.), applications that can take advantage of dual processors, or just want to have the baddest, meanest Mac on the block, the top of the line Professional Dual G4/1GHz Tower may be overkill for your needs.

If you want a machine that is forward looking, this also may be a reason to get the dual processor machine. It will definitely give you more freedom and flexibility than the 933 MHz machine one or two years down the road, as more and more applications get coded to take better advantage of its power.

It may also hold its resale value better when it comes time to trade it in for the next "biggest, baddest machine" in the crowd.

As you will see from the results below, in some Photoshop work and other applications that pour code into dual processors, the Dual GHz machine really shines, turning in a performance improvement of from 50-90%. However on many other applications the performance improvement is more modest, more or less in step with rising clock speed.

Hardcore Gamers have one choice upgrade their graphics cards or upgrade to a new Tower machine. Both the Dual Gigahertz machine and the 933 turn in frame rates that are double that of the last generation of Towers .... at least when running Quake in quality mode. Our Quake III scores showed little difference in performance between the new dual processor machine and the 933. This may change when/if games are written to take advantage of dual processors. (We are working on getting a suite of games together that will give a well rounded performance picture of new machines ... if you have any suggestions let us know).

Additional Resources

 

Check prices at:

 

Now without further ado ... on to the data!

Difference and similarities in processor and memory systems of each machine


  Power Mac G4/733 Power Mac G4/867 Power Mac Dual G4/800 Power Mac G4/933 Power Mac Dual G4/1GHz
Processors G4/733 G4/867 2 x G4/800 G4/933 2 x G4/1000
L2 Cache 256k @ 733 MHz 256k @ 867 MHz 256k @ 800 MHz 256k @ 933 MHz 256k @ 1 GHz
L3 Cache none 2 MB @ 217 MHz 2 MB @ 200 MHz 2 MB @ 233 MHz DDR 2 MB @ 250 MHz DDR
Memory 768 MB PC 133 768 MB PC 133 768 MB PC 133 768 MB PC 133 1 GB PC 133
Graphics Card NVIDIA GeForce2 MX with 32MB of SDRAM in 4x AGP slot. NVIDIA GeForce2 MX with 32MB of SDRAM in 4x AGP slot NVIDIA GeForce2 MX with TwinView and 64MB of SDRAM in 4x AGP slot. NVIDIA GeForce4 MX with 64MB of DDR SDRAM NVIDIA GeForce4 MX with 64MB of DDR SDRAM
Drive 40 GB @ 5400 rpm 60 GB @ 7200 rpm 80 GB @ 7200 rpm 60 GB @ 7200 rpm 80 GB @ 7200 rpm
Operating System 9.2 9.2 9.2 9.2.2 9.2.2
Price $1,699 $2,499 $3,499 $2,299 $2,999


"Real World" Tests

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 Quake III & Cinebench 2000 tests) were timed with a stopwatch. The times are then converted to percentages relative to the Power Mac G4/933, which is set to 100%. For all scores, higher numbers are better.

Drive & Finder Tests



The problem with doing these kind of comparisons is that, when you get an odd result (like for the 867 machine above), you would like to go back and retest to confirm your results. Unfortunately by that time the machine is long gone. Take the 867 result with a grain of salt...

This looks about right. Both the 933 & Dual GHz have drives that spin at the same speed, as do the 867 and Dual 800. The 733 spins at a slower speed

All the machines except for the 733, write CDs at 8X. The 733 writes CDs at 12X. It will be interesting to see how the new single processor 800 MHz machine stacks up as it writes CDs at 24X.

AppleWorks 6 Test

Here the machine without a L3 cache pays the price. You should see something similar with the new 800 MHz Tower. This test is a search and replace in a very large document.

Quake III Tests

Processor speed is king when running Quake at fastest mode ... but who wants to do that

Amazing performance for the new machines. The GHz machine turns in 115 fps and the 933 111 fps. This is due almost totally to the graphics card .. not to increased processor speed. We are looking forward to testing the new 800 MHz Tower, which has a different graphics card, to see how it does in comparison.

Photoshop 6 & Other Data Crunching Tests

You can see again that the 733 machine is hurt by it's slower drive, and perhaps by the lack of the L3 cache

.This is one application where dual processors can make a difference.

No dual processor advantage here. Not all of the functions of Photoshop utilize dual processors. Nor are they written to take advantage of the extra performance power of the G4. It will be interesting to see what the OS X version of Photoshop will bring.

This is a raw processing power test

Ok, this is what it is like when a program has been well written to take advantage of dual processors

Encoding/Decoding Tests

The two new machines are pretty evenly matched. Why? Only one of the processors on the GHz machine is being utilized

Raw processing power is what is going on here. Again the dual processor machine is under utilized by the application


Actual Scores - In seconds except for the Quake III scores and the CineBench 200 scores

Machine Copy 60 MB Folder Copy 128 MB File AppleWorks 6.2 Scroll Test AppleWorks 6.2 Search & Replace Test Time To Analyze 2 Million Lines Of Code And Generate Report Quake III Fastest Setting
Power Mac Dual G4/1GHz 10.2 8.41 30.61 6.47 20.31 130 fps
Power Mac G4/933 10.33 8.03 33.88 6.91 21.5 124.4 fps
Power Mac Dual G4/800 10.71 9.39 31.11 7.67 24.47 100.8 fps
Power Mac G4/867 14.92 9.11 28.29 7.43 22.28 112 fps
Power Mac G4/733 12.7 11.28 37.7 18.21 28.41 71.5 fps
 
Machine Quake III High Quality Setting CineBench 2000 - RayTracing CineBench 2000 - Shading iTunes 1.1 MP3 Encode QuickTime 5 Sorenson Encode SuffIt 6 Expander
Power Mac Dual G4/1GHz 115.3 fps 21.77 points 10.04 points 27.14 192.18 9.23
Power Mac G4/933 111.1 fps 11.69 points 9.56 points 28.05 195.33 9.95
Power Mac Dual G4/800 56.4 fps 17.02 points 7.78 points 40.65 239.66 11.66
Power Mac G4/867 54.3 fps 10.76 points 8.7 points 39.08 211.02 11.22
Power Mac G4/733 53 fps 8.71 points 7.45 points 37.88 282 13.99
 
Machine Photoshop Launch Photoshop 6 - AltiVec Tests Photoshop 6 - Non-AltiVec Tests Re-Boot Test Burn CD  
Power Mac Dual G4/1GHz 6.43 18.98 101.60 29 122.2  
Power Mac G4/933 6.29 27.49 108.78 29.28 119.98  
Power Mac Dual G4/800 6.75 24.31131.31 41.14 127.95    
Power Mac G4/867 6.87 29.89116.82 34 128.13    
Power Mac G4/733 9.19 38.84 161.66 39.37 90.02