From the source code, joml.Math.sin(float) will always use java Math.sin() without an optional table lookup method. But sin_theagentd_lookup uses floats, it's okay to apply the method to sin(float) as an option. Is this intentionally disabled? If so, should I replace Matrix4f with Matrix4d if I need fast math?
public static float sin(float rad) {
return (float) java.lang.Math.sin(rad);
}
public static double sin(double rad) {
if (Options.FASTMATH) {
if (Options.SIN_LOOKUP)
return sin_theagentd_lookup((float) rad);
return sin_roquen_newk(rad);
}
return java.lang.Math.sin(rad);
}
Thanks for reporting.
You're right, the FASTMATH path can of course also be used for the float-overload.
Is in the 1.10.4-SNAPSHOT now, will come with the 1.10.4 release.
From the source code,
joml.Math.sin(float)
will always use java Math.sin() without an optional table lookup method. Butsin_theagentd_lookup
uses floats, it's okay to apply the method to sin(float) as an option. Is this intentionally disabled? If so, should I replace Matrix4f with Matrix4d if I need fast math?