In the pixel shader i can do one of this two configuration
--reading from a R32G32B32A32 texture one time using Load() one time
--reading from R32 texture four times using Load() four times
The first must be faster, lets say the pixel shader runs around 1,000,000 times is there a big difference between the two?