OTOY Forums

Hi,

I read somewhere that the 4xx series was deliberately crippled so that DP calculations run at 1/5 the speed of single precision floats, while the Fermi architecture can manage 1:2 DP vs SP.

I never noticed this before but when you search for it there are several sites which indeed mention it. Is this really true and if so does this hamper Octane?
Can´t remember seeing benchmarks here where Octane speeds up enormously on the equivalent profi cards?

Max.

The "crippling" only happens with standard DX/OGL apps. It doesn't affect CUDA (and hence Octane Render) at all.

Regards
/M

Thanks a lot. I was convinced it couldn´t be true but some people kept repeating it.