Accelerate Your Mac! - 3dfx Mac Voodoo5 5500 PCI Review

MacBench 5.0 Tests:
I ran MacBench 5.0 Graphics tests at 1024x768, thousands colors and Publishing Graphics tests at 1280x1024, millions colors (the Rage128 PRO AGP would not show 1152x870 in the list with my monitor). 2d performance was one area where the previous Voodoo3 usually fell flat. Most all modern graphics cards are capable of far higher resolutions and refresh rates (see the specifications page for maximum resolutions of the Voodoo5).

The MacBench Graphics and Pub. Graphics tests run a scripted series of Quickdraw calls that simulate actual applications like Photoshop, Word, Pagemaker, etc. running (in fact you can see the applications emulation running during the tests). All tests with the Voodoo5 were with the default font cache and picture cache enabled. In the future when time allows I want to test with these caches disabled (perhaps when the final drivers are released).

[Note - readers say 2D mode only uses 1 chip, which is why it was faster by a hair in Macbench I assume (there is some overhead perhaps from dual chip mode).]

Applications Scrolling Tests:

I also ran scrolling tests in Photoshop and Word98. Remember these are 'stopwatch' tests so there is some margin for human error in starting and stopping the clock (consider scores within a fraction of a second as about equal due to the margin of error). Also remember it's hard to 'see' a small performance difference in actual use. Photoshop was allocated 10x the 18MB image size and Word 98 was allocated 60MB to minimize any swapfile activity. (Photoshop efficiency was at 100%.)

The Photoshop test image was a 18MB disk file size photo at 300DPI. I zoomed to 400% and moved the scrollbars to the top and left edge. I then measured the time to scroll completely down and to the right.

For Word 98 scrolling tests I used a document with more than 2,000 pages of mixed font sizes. I first measured the time to scroll to the end of the documented and then back up to the top.

2D Graphics Performance Summary:
Even with some dual processor overhead, the higher MacBench scores with single chip mode vs the default 'faster' (dual chip mode) don't make sense. Granted 2% is a tiny difference, but if both chips do work together in 2D I'd not expect that mode to be slower. It's unknown if the current state of the Voodoo5 drivers affects its 2D performance as well as 3D. Although the Rage128 PRO AGP had a slightly higher MacBench Pub Graphics score, remember it's running in a 66MHz AGP slot and the Voodoo5 was running in the G4/500's 33Mhz slot. Since the current Voodoo5 driver has some known performance issues, it will be interesting to see how later drivers perform on the Voodoo5.


Review: 3dfx Mac Voodoo5 5500 PCI Graphics Card Review date: 7/26/2000
2D Performance Tests Results
Intro \| 2D Performance \| 3D Performance \| Game Performance \| Movie Playback \| Controls/Design \| Summary
MacBench 5.0, Photoshop and Word98 Performance
MacBench 5.0 Tests: I ran MacBench 5.0 Graphics tests at 1024x768, thousands colors and Publishing Graphics tests at 1280x1024, millions colors (the Rage128 PRO AGP would not show 1152x870 in the list with my monitor). 2d performance was one area where the previous Voodoo3 usually fell flat. Most all modern graphics cards are capable of far higher resolutions and refresh rates (see the specifications page for maximum resolutions of the Voodoo5). The MacBench Graphics and Pub. Graphics tests run a scripted series of Quickdraw calls that simulate actual applications like Photoshop, Word, Pagemaker, etc. running (in fact you can see the applications emulation running during the tests). All tests with the Voodoo5 were with the default font cache and picture cache enabled. In the future when time allows I want to test with these caches disabled (perhaps when the final drivers are released). [Note - readers say 2D mode only uses 1 chip, which is why it was faster by a hair in Macbench I assume (there is some overhead perhaps from dual chip mode).] Applications Scrolling Tests: I also ran scrolling tests in Photoshop and Word98. Remember these are 'stopwatch' tests so there is some margin for human error in starting and stopping the clock (consider scores within a fraction of a second as about equal due to the margin of error). Also remember it's hard to 'see' a small performance difference in actual use. Photoshop was allocated 10x the 18MB image size and Word 98 was allocated 60MB to minimize any swapfile activity. (Photoshop efficiency was at 100%.) The Photoshop test image was a 18MB disk file size photo at 300DPI. I zoomed to 400% and moved the scrollbars to the top and left edge. I then measured the time to scroll completely down and to the right. For Word 98 scrolling tests I used a document with more than 2,000 pages of mixed font sizes. I first measured the time to scroll to the end of the documented and then back up to the top. 2D Graphics Performance Summary: Even with some dual processor overhead, the higher MacBench scores with single chip mode vs the default 'faster' (dual chip mode) don't make sense. Granted 2% is a tiny difference, but if both chips do work together in 2D I'd not expect that mode to be slower. It's unknown if the current state of the Voodoo5 drivers affects its 2D performance as well as 3D. Although the Rage128 PRO AGP had a slightly higher MacBench Pub Graphics score, remember it's running in a 66MHz AGP slot and the Voodoo5 was running in the G4/500's 33Mhz slot. Since the current Voodoo5 driver has some known performance issues, it will be interesting to see how later drivers perform on the Voodoo5. The next page covers 3D performance in non-game applications.

Index of 3dfx Voodoo5 5500 PCI Review Intro \| 2D Performance \| 3D Performance \| Game Performance \| Movie Playback \| Controls/Design \| Summary - or - Back to www.XLR8YOURMAC.com
Copyright © 2000, all rights reserved. All brand or product names mentioned here are properties of their respective companies. Site terms and conditions of use.