-snip-
Perhaps there is an implementation in C++ or Python code?
A negative point is the same point only with a negative coordinate Y
In python i use operator.neg(point_y))%p then ordinary addplt
p in your case is modulo of the curve.
To be honest operation
(p-y) is faster than
(-y)%p/ I have just tested for random 100 million 256bit numbers just the subtraction (p-y) and division by modulo (-y)%p --> division by modulo took 1.4527 longer