|F4 Benchmarks: Voodoo3
by Bubba "Masterfung" Wolford and Pierre "Papadoc" Legrand
A new era is upon us. The 3d wars are heating up again with the introduction of 3dfx's newest chipset, labeled "Voodoo3."
Like most of you, I have been following the news on Voodoo3 for weeks now. As sites on the 'net began posting benchmarks I knew we flightsim buffs were in for some heavy debate until we could get our hands on the newest chips and test them in our "world."
Benchmark Standards and Combat Sims
When I check a site like Toms Hardware or Sharky Extreme looking for information, I always know I am going to find interesting info, but I know that it will be limited in relevance to the games I am playing. Since I do not play Quake 2, Shogo, or any other benchmarking first person shooter (FPS) enough to care about their FPS numbers, I am always out of luck and extremely frustrated when trying to find data relating directly to my games.
At least I know I'm not alone! My feelings are shared by many of you who email us here at COMBATSIM.COM™ looking for information on combat sims benchmarked using newer video cards like Voodoo3 or TNT2.
Not too long ago I was searching for information about 3dfx's last chipset, Banshee. I found some good data about Banshee on Tom's Hardware. Tom's opinion of the Banshee was that the Banshee was not a real competitor in the 3d arena.
Subsequent events have made Tom's results questionable. Why did he take the position he did? The Banshee lacked a 2nd Texture Memory Unit (TMU), and could not be linked with a second Banshee to push SLI like it's older brother V2, and thus, why buy Banshee when one could get a better one card solution like TNT?
Well, Tom's opinions are often well founded and I respect them, but unfortunately he bases ALL his video card assessments on which cards run Quake and other 3d shooters the best. This just doesn't apply in the miliary sim world. Our main concerns revolve around CPU, the absolute speed (MHz) of the memory and core clock rate of the card.
The Squeaky Wheel...
One of the reasons it is so difficult to get data on games like Falcon 4.0 is summed up in one word: ORGANIZATION. Simply put, we are not organized as a group. The reason Quake and Quake 2 became such a reliable and often cited benchmark is because they have become an industry standard. Sure Quake and Quake 2 are very popular games, but one of the reasons so may people play them is because they want to see how their computers stack up against the next guy's machine and both of those games offer an effective way to measure.
In the flightsim arena we desperately need a reference point to put ourselves on the map as a credible and important market. We need a benchmark that can give us information relevant to what we play. This in turn will enable us to stop basing our monetary decisions on games that are not relevant to our applications. Until recently, not one such benchmark existed and even now I can count the number of simulations that post FPS numbers on one hand.
The Irony of Current Benchmarks
Both of the above problems lead to our third issue. Since we are not always seen as an organized or sizable group with singificant and unique hardware needs, we get overlooked. Consequently, its tougher to get hardware to actually do the necessary benchmarks.
This results in a laughable situation. When Voodoo3 benchmarks first appeared a few weeks ago, every website that had a reference board was showing the SAME benchmarks with the SAME games and the SAME frames-per-second (FPS) numbers. After reading my second Voodoo3 3500 review, I was no better informed, and not very happy. Nothing any of the websites showed gave me (or any of you) any information that we wanted to see.
Ten websites got early Voodoo3 3500 boards and ALL of them were displaying the same games and FPS numbers! What is the point? Why did the hardware distributors even bother? And while all these sites were showing the same irrelevant results, it was impossible to find any numbers on the net that were relevant to our hobby.
Profile of a Sim Fan
Recently on the flightsim newsgroup there was a survey asking everyone to post their age. I am 27 and I can count on one hand those that were younger than me among the hundreds that posted their stats.
Based on the unofficial survey I would guess the median age of sim fans is 30+. I would also wager we tend to have a higher median salary than those who play Quake 2 and the like, which follows from our median age and also the intellectual intensity of the games we choose.
So what does this mean? It means that military sim fans are devoted gamers with a disporportionate amount of purchase power who run the most demanding applications out there. We are struggling to get a consistent 20 fps in some of the games we run, where the average Quake player is already over the edge at 40 to 60 fps.
Because the games we play are so demanding, sim fans must upgrade and purchase higher end equipment more often in order to pursue their hobby. This year alone, I have built two Pentium II 400-450 Mhz machines each with 256 Megs of RAM mostly for playing and testing Falcon 4.0.
Go to Page Two
Copyright © 1997 - 2000 COMBATSIM.COM, INC. All Rights Reserved.
Last Updated April 12th, 1999