The precision of math operations performed by the shader.
The set of possible values for the precisionHint property is defined
by the constants in the ShaderPrecision class.
The default value is ShaderPrecision.FULL. Setting the precision to
ShaderPrecision.FAST can speed up math operations at the expense of
precision.
Full precision mode (ShaderPrecision.FULL) computes all math
operations to the full width of the IEEE 32-bit floating standard and
provides consistent behavior on all platforms. In this mode, some math
operations such as trigonometric and exponential functions can be
slow.
Fast precision mode (ShaderPrecision.FAST) is designed for maximum
performance but does not work consistently on different platforms and
individual CPU configurations. In many cases, this level of precision
is sufficient to create graphic effects without visible artifacts.
The precision mode selection affects the following shader operations.
These operations are faster on an Intel processor with the SSE
instruction set:
sin(x)cos(x)tan(x)asin(x)acos(x)atan(x)atan(x, y)exp(x)exp2(x)log(x)log2(x)pow(x, y)reciprocal(x)sqrt(x)