They'd have to either sacrifice yield or sacrifice FP32 units.

Also, if they put FP64 support in their gaming cards they wouldn't be able to charge scientists and engineers extra.

> Also, if they put FP64 support in their gaming cards they wouldn't be able to charge scientists and engineers extra.

Bingo. Not many entertainment problems demand such precision. Why not charge more for it?

Well, if AMD had taken the need to get their GPGPU offering out the door seriously or if they had later taken the need for CUDA compatibility seriously then Nvidia would have had to compete.

Alas, we're talking about a company that doesn't even think it's important for their installer to reliably replace existing drivers. The result: I get to pay through the nose for my predecessor's CUDA lessons. Yay.

Wat. They don't make gaming cards nor scientific cards. They make GPUs that go into iPhones and iPads.

Who is going to use FP64 in iPads and Android tablets?