Does it mean that when doing a conversion from sRGB encoding to physical intensity encoding we have to extend the number of bits to encode the physical intensity values to avoid rounding errors in the sRGB encoding ?

I guess that that the required number of bits to encode physical intensity values depends on the operations. performed. The author suggest using floats, but this means 3x4 bytes and 4x4 bytes with the alpha channel. Would 16 bit unsigned integer be enough ? Floats are ok when using graphic cards, but not ok when using the processor.