The Source for Mac Performance News and Reviews

Review: Powerlogix's 1GHz G4 Upgrade for the Cube Tests vs Original G4/450 and Dual G4/500 By Mike Published: 8/19/2002 (Updated 8/21 for RTCW MP timedemo tests)

3D Game Performance Tests

Intro | Benchmarks  | Apps Tests  |  Game Tests | Installation | Specs/Design

Game Performance Tests

Note: This review used the 1MB L3 cache version. New models have 2MB of L3 cache in the 1GHz and 800MHz models. This page lists test results in popular 3D games like Quake3, Unreal Tournament, Return to Castle Wolfenstein and Medal of Honor Allied Assault (MOHAA).

Quake3 Tests:

Tests were done with the latest Quake3 version (1.31b5) as of this review date. I used the default "High Quality" Quake3 setting w/all game options on. (Only the resolution was changed.) The Quake3 config file had only one tweak - s_chunksize was set to 4096 vs the default 512. For the dual 500 test, r_smp was set to "1" (to use both CPUs). The single 500 results were with r_smp=0. I expected better performance from the 1 GHz upgrade (i.e. closer to the dual 500's results.) In games that don't support dual processors (Return to Castle Wolfenstein, MOAHH, etc.) the 1 GHz upgrade did perform better than the dual 500. (RTCW was noticeably better with the 1GHz during play than the dual 500.)

Unreal Tournament Tests:

I used the latest UT versions as of this review date (v436 in OS 9 tests, PR3 for OS X tests). Medium Quality settings, music volume set to 0, decals and dynamic lighting enabled, "Min Desired FPS" set to 0. (Rave Mode used for OS 9, since it runs better than OpenGL mode even on Nvidia cards in my experience. OS X used OpenGL mode.) OEM GeForce3 card used for all tests. I used UTBench rather than Wicked400, since UTBench has more varied gameplay, but still has lots of action. Framerates seen in actual play will be higher than what is shown with this benchmark however (in my experience). During actual play with the 1GHz upgrade for instance I usually saw 40-60FPS rates, but I can't watch the instaneous FPS meter constantly during gameplay (and stay alive at least).

Return to Castle Wolfenstein Tests:

RTCW doesn't work with config tweaks like "r_smp" for Quake3 to use both CPUs of dual G4 systems. (Just like MOHAA, hopefully that will be fixed in a future update. I have not read where v1.33 fixes the problems with SMP in RTCW.) 8/21/2002 Update: Reader Tzunami sent a Multiplayer demo to use for performance tests. (You can download the demo using the button in the sidebar at http://ultimatedataserver.de.vu/.) I've run tests with the original G4/450 and the 1GHz PL upgrade with this demo in both OS 9.2.2 and OS X 10.1.5. (I haven't tested this with the DP500, since it requires more involved heatsink mods and kernel paniced when I tried to run RTCW again recently (this time with a better fitting home-made heatsink, but one that didn't cool as well apparently as for the first time I saw a KP with it. and the core was very hot when I removed it.) I didn't have a timedemo to run when I first posted this, so I used a saved game from mission 1 (inside the castle area) where I had seen very disappointing performance in play before. With draw fps enabled, I usually saw under 20FPS during play in this area - during firefights even with only one or two enemies firing, FPS often dipped to 14 frames per second. The 1GHz upgrade made the biggest difference in this game, with FPS rates going from 14-20 typically with the other CPUs to 22-40 (or higher depending on the environment). Granted that seems low to PC users or more modern Macs, but regardless the 1GHz upgrade did make it much more playable. Here's a list of FPS ranges I saw during play in mission 1 with the each of the 3 CPUs installed in the Cube. (The start of the saved game is in the room where there are holes in the floor planks with two german solders below.) Game settings were 640x480/32bit, medium detail, no Fog. Mission 1 Play: G4/450: 14-20 FPS (almost always under 20) DP 500: 14-21 FPS (often under 20 during firefights) PL 1GHz: 22-40 FPS (usually in the upper 20s to 30s) I also timed how long it took to load the game and a saved mission with each CPU: Load Game G4/450: 15.6 sec DP 500: 15.1 sec PL 1GHz: 14.9 sec Load Saved Game G4/450: 46 sec DP 500: 37 sec PL 1GHz: 30.8 sec I used RTCW v1.32. (v1.33 was released after testing had been done on two of the CPUs. Some people have also reported problems/corruption with the 1.33 update, but I've not been able to download it yet.)

Medal of Honor Allied Assault Tests:

I enabled framerate display in MOHAA (console "FPS 1" command) and recorded the typical framerate shown during firing the mounted Machine Gun in the training mission (while destroying the target). The fire button was held down for continious firing. (Framerates increased a bit after the target was destroyed but the figures below were taken during the target breaking up.) Game settings were 1024x768/32bit mode, medium settings. I used the latest MOHAA version (1.12) as of the review date. (Like the RTCW tests, I wish I had a standard timedemo or benchmark for this game, but this is the best I can do for now.) MG Firing FPS (OS X) G4/450: 20-23 FPS DP 500: 30-33 FPS PL 1GHz: 31-34 FPS I also timed how long it took to load the saved training mission with each CPU: Load Saved Game (OS X) G4/450: 27.4 sec DP 500: 23 sec PL 1GHz: 18.2 sec OS 9.2.2 results follow: MG Firing FPS (OS 9) G4/450: 27-29 FPS DP 500: 29-31 FPS PL 1GHz: 39-42 FPS Load times increased in OS 9: Load Saved Game (OS 9) G4/450: 42.9 sec DP 500: 37.9 sec PL 1GHz: 31 sec

Other Sources of Game Performance Comparisons: The Video card reviews page lists articles and reviews with comparisons of literally every fast Mac 3D video card currently on the market. The Systems page includes reviews of various G4, G3 and PowerBooks. Usually those reviews include game performance tests. The searchable Mac Game Framerates database has thousands of entries of Mac 3D game framerate reports using games like Unreal, UT, Quake1, Quake2, Quake3, etc., as well as other benchmarks. Systems range from 603e models to Apple's latest AGP/G4s. You can limit searches by game title, video card, CPU speed, Mac model (or any combination). This database can be a great tool to see what effect CPU speed and video card have on game performance.

The next page comments on installation of Powerlogix CPU upgrade in the Cube. Or you may use the links below to jump to a specific page.

Index of PL 1GHz Upgrade Review Pages Intro | Benchmarks  | Apps Tests  |  Game Tests | Installation | Specs/Design - or - Back to WWW.XLR8YOURMAC.COM

Copyright © 2002. No part of this sites content or images are to be reproduced or distributed in any form without written permission. All brand or product names mentioned here are properties of their respective companies. Users of the web site must read and are bound by the terms and conditions of use.