Max Payne Benchmarking

20th August, 2001 / by Leonidas / translated by RAL & Houston / Page 6 of 6

      Final thoughts

A few words about the minimal fps: Sometimes the game shows the right value, sometimes it shows nothing but "0:00". We suggest you to retry it again and again until it works :-).

It also should be noted that average value is resulted from thousands of frames, where the minimum value is only one single value. Every statistician will tell you not to establish any conclusion without regarding enough sample values. To judge a video card solely on the minimum value would be as ridiculous as to ask one single citizen to choose our president.

To come back to the four demo sequences, my personal favorite is the first demo "PCGH´s Final Scene No.1". The second is a bit too short in time; the third shows too less action on the screen and the fourth provides too high average values not representatively enough in relation to the rest of the game. One of the important points of a good benchmarking sequence is to represent the whole game in its average fps values. The first demo scene fits this requirement at best where the others are slightly worse.

I believe that it is not enough to create a benchmark scene only to provide a comparison between different video cards. The reader expects an overall meaningful fps value that is realistic regarding to real in-game framerate. But that´s only my personal opinion on what a good benchmarking scene should provide. So we have included four different kinds of demos so that everyone may have their own favorite(s).

Finally I would like to thank all the people involved in the creation of this mod:
• SirElvis for the programming of the mod and his very patience to fit my always changing wishes of the mod´s behavior.
• Turrican webmasters from the sites Max Payne Zone and 3DActionZone for providing the contact with SirElvis.
• Thilo Bayer from PC Games Hardware for the first and second benchmark demo and for his many useful tips.
• Raphael auf der Maur from 3D Concept for providing the third benchmarking demo.
• RAL from 3DCenter & Houston from Hard Tecs 4U for the translation of this article :-).

 

Update November 9, 2001:

"The quest for anisotropic filtering in May Payne" - a reply from Markus Stein, member of the Max Payne development team:

If you enable anisotropic filtering, the game uses the highest available quality level of anisotropic filtering. Benchmarking with enabled anisotropic filtering is problematic.

We recommend to disable this option in favor of getting comparable benchmarking results. If you really need benchmarks with anisotropic filtering, please enable this feature from the video card driver menu.

Update June 25, 2002:

It is recommend to use a noCD patch for Max Payne benchmarking. Benchmarks with noCD are 15 - 25 percent higher than benchmarks without noCD patch.