A friend who runs a very cool display company (true 3D volumetric displays, not spinning things), asked me to do a quick benchmark for him. As I was freshly done with some other GLperf stuff, I agreed. FWIW getting GLperf to compile on late model Linux is … well… interesting. Especially the 64bit versions. Lots of hardwired bits in the build. Annoying.
His test was fairly simple, but it hit the critical points he needed to hit. This is important, as this is where his code spends most of its time. That is the microbenchmark in this case was meaningful to him.
The timings that we observed were for 3 different graphics cards using sets of operations that he does all the time for his hardware. He has been convinced over time that the commodity graphics cards are identical to the professional cards. I thought otherwise, and proposed the test, as he had been running out of steam on the commodity setup.
First test was a simple frame rate test, as they need to toss lots of information down the GL pipeline as quickly as possible.
GEforce4 440 MX: 529.400 FPS
Quadro FX 1400 : 6759.200 FPS
Quadro FX 4500 : 12134.200 FPS
Similar in many aspects to what their code does at one point. Next test was a 3D texture test.
GEforce4 440 MX: 62.400 FPS
Quadro FX 1400 : 271.000 FPS
Quadro FX 4500 : 515.600 FPS
I can’t release this code, or talk much about it. Interesting though.