Hi,
Is the AMD implementation supports float32, float64?
If yes, how to access elements?
example: float16 elements can be accessed using, s0 - sf
AFAIK, vector types support number of components: 2, 3, 4, 8, and 16.