Are there Cg-like vector operators such as operator+(float3,float3), operator*(float3,float3), etc, defined somewhere in the api, or were these intentionally left out?
You can overload operators yourself, but they aren’t included by default.
Unlike many previous GPUs, the G80 ALUs are scalar, so there’s no inherent need for vector types in order to get high performance.
John Stone …also at UIUC :)