280 GTX (g200) vs 285 GTX (g200b)

I am looking into creating a budget CUDA setup, and I was comparing the differences between the 280 GTX and the 285 GTX. It seems that in the process of the 65 nm to 55 nm transition, I read that the FP64 core was removed? Per http://forum.xcpus.com/nvidia/15259-foldin…nformation.html

This post was dated way back in January, and I was wondering if it holds true today? Am I better off having a 280 GTX for my budget CUDA needs or is it okay to have the 285 GTX?

Thanks for reading, and I hope this is the right place to be posting about this.

eh? no, they’re functionally identical (both are compute 1.3, support DP, etc.)

The only difference between the GTX 280 and GTX 285 is the power draw. GTX 285 draws 183W, and is powered by two of the 6-pin PCI-Express power plugs. GTX 280 draws 236W, and is powered by a 6-pin and an 8-pin PCI-Express power plug. Some power supplies don’t have the 8-pin plug, so I would investigate that issue first.

Core and memory clock rates are also higher on the GTX 285. I see ~10% better performance on the 285.