Ever wondered which card is better out of these two beasts? Well we've found a guy who tested these monsters to the max.

Canadian Eric St.Martin of Ubermedia is the man in question and here's a brief account of his story:

Eric prepared a Motion file with everything required to push the GPUs of both cards to the limit.  The main idea was to test the real-time playback capacity of the cards with enough room to quantify the performance of each card properly.

Eric created a 5 seconds animation in 1080p @ 60fps.  It’s a custom setting but he needed to test the performance of the cards beyond 30fps.  He also ran some preliminary tests and the FPS in the project settings have no impact on the performance of the real-time playback.  That is to say that if the card outputs 18fps playing a given project, this number of FPS played back will be the same regardless of the target playback rate of the project.  The card spits out everything it can up to the playback rate of the project.

When we playback an animation project in real-time in Motion, there are many options available impacting the render quality.  Obviously, the more options you activate, the less fluid the playback.  So Eric tested with a variety of configuration of the following options - Motion Blur, Reflections, Lights, Shadows. Using all three render precision settings.

Finally, he flushed the memory before each test (Mark->RAM Preview->Clear RAM Preview), switched on “loop playback” and the FPSs were noted for the first playback pass and then, the second.  Indeed, Motion seems to cache some information in RAM during the first playback pass and when the playback loops for the second pass, it’s almost always faster.

The result:  The ATI 5870 won.

When you consider its half the price of the Nvidia Quadro 4000, but performs better/quicker with Motion. Definitely worth contemplating on your next purchase.

Circumstance forced Eric St.Martin into purchasing both cards. You can read the full article about his triumphs and tribulations here. There's also a table of all the results.