Open thiagomanel opened 5 years ago
Analyzing the largest pixel difference between the cplusplus and r code outputs, the result was problems from the 10 ^ -7 scale, because the output pixels in R are only accurate to 10 ^ -6. There was only one update in the constant PI for better accuracy resolved in #20.
It seems we have problems at 10^-3 scale