Stupid question, probably, but, why does
float2 charSizef = (float2)(charSize.x, charSize.y);
work, whereas
float2 charSizef = (float2)charSize;
Doesn’t, where charSize is an int2 input to my kernel function.
Also, what’s the performance-hit from casting between int/int2 and float/float2?
a|x