Thilo Bayer, hardware editor at PC Games Germany has received an GeForce FX for testing. The print magazine including the test will be on sale in the first week of February (February 5th, 2003 to be exact). More interestingly he has posted some early findings in the 3DCenter forums. I will try to summarize his findings.

Read more...

Preliminary Benchmark: Test system: AMD Athlon XP 2600+ 512MB RAM nForce 2 motherboard UT2k3 2166 - Flyby (2X FSAA / 4:1 Anisotropic Filtering 1024x768x32Bit)ATi Radeon 9700 Pro - 110 FPSNvidia GeForce FX - 143 FPSNvidia Ti4600 - 65 FPS
UT2k3 2166 - Flyby (1024x768x32Bit)ATi Radeon 9700 Pro - 201 FPSNvidia GeForce FX - 196 FPSNvidia Ti4600 - 170 FPS UT2k3 2166 - Botmatch (2X FSAA / 4:1 Anisotropic Filtering 1024x768x32Bit)ATi Radeon 9700 Pro - 65 FPSNvidia GeForce FX - 65 FPSNvidia Ti4600 - 65 FPS
UT2k3 2166 - Botmatch (1024x768x32Bit)ATi Radeon 9700 Pro - 71 FPSNvidia GeForce FX - 68 FPSNvidia Ti4600 - 71 FPS According to Thilo Bayer his sample board works with some basic hardware monitoring. The GeForce FX is adapting to the working environment you are currently running it in. This means: 1.) Working in 2D mode (a.k.a. desktop / office): core clock = 300Mhz / memory clock = 600Mhz 2.) Working in 3D mode (a.ka. games): core clock = 500Mhz / memory clock = 1000Mhz Mr. Bayer describes the FXFlow cooling system to be annoyingly loud and recommends 'ear plugs', which was to be expected. Furthermore his writings state that the A2 stepping is the final retail revision of the GeForce FX GPU/VPU. His card is apparantly no sample but a production card. He promised more results for tomorrow.