Darkman wrote:got a Creative Annihilator Pro 32MB card (Geforce256 DDR) for £1 , and decided to try it out for a bit in the same Athlon 1400 m […]
Show full quote
got a Creative Annihilator Pro 32MB card (Geforce256 DDR) for £1 , and decided to try it out for a bit in the same Athlon 1400 machine as above , to see how it competes with the Voodoo 5500.
Drivers used were 30.82
<image snip>
I would have to assume the higher score is probably due to the DX7 functionality the Geforce256 has, when I tested Quake3 , the timedemo got 64fps compared to 77 on the V5500 at the same settings, Nevermind Glide games like UT99 where the V5500 got 74 in the cityintro demo , while the GF256 DDR got 40.5 at the same res & detail level in D3D . Even in Max Payne the GF256 was about 20fps slower when using hardware T&L.
have to wonder why the benchmarks at various review sites showed the GF256 DDR beating the Voodoo in most situations, Weaker CPUs maybe? bad drivers?
It's been a common practice for a great number of years that manufacturers send specialized versions of drivers along with their review samples to various test labs around the world when testing their new cards for displaying comparisons on the internet. They hand-tweak these drivers to make the various cards run faster in certain tests so people looking to these websites will think their cards are faster and then go buy them. Even when the companies know that the retail product with retail drivers will not perform anywhere near as well as they did in the versions sent to test labs. This is still done today even with today's modern video cards.
The "GOOD" test labs will throw out included drivers and use retail drivers off the websites at the time of testing. Not all test folks will do this. I don't know exactly how it works, perhaps maybe some sites have to sign legal agreements to use the provided drivers.. I don't really know. But it's been common place in the industry for years. Most people know better than to follow review sites and buy based on their results.
The one I usually refer to for performance is "Passmark Performance Test", which is a private run thing where individuals on their own computers test their own cards (presumably with reference drivers), and then we can find their submissions via online database in the program and add them together. And also everyone runs the exact same program, and the same test. We can also see cpu and hardware they used for testing.. and then match it up so all tests have the same hardware at the same speed, etc. I haven't looked before, but I might go check if they have voodoo5 and older cards in there.
Alternatively we can use their website where they display an overall average of all global submissions and perform a weighted average and then graph it. So it's not swayed by one individual result, and has a rank of all cards.
EDIT: I went and looked inside Passmark Performance Test and the baselines that let us search online for other submissions, there's people that have put in submissions back to the Pentium III and Geforce FX series. Also some AMD AthlonXP and Duron results are in there. Folks can just use that and if you "select baseline" it puts it in a nifty colored graph you can look at, and run it on your current system and then compare how yours does verses others.