loss-of-significance is a phenomenon occuring in numerical programs when the result is significantly less accurate than the resolution of the floating point numbers used. Often, but not always, loss-of-significance can be circumvented by making the required adjustments to the numerical algorithm.
Questions tagged [loss-of-significance]
2 questions
2
votes
1 answer
Java - Comparable interface with very small number
I have this class that stores the latitude and the longitude of a coordinate and it distances from the center.
At some point in my algorithm I have a list of these Coordinate and I want to sort them by their distance from the center. Thus, I…
ale666
- 41
- 4
-1
votes
1 answer
Discrepancies in C++ float subtraction
I'm working on a Vec3 template class that (not surprisingly) stores a vector of three elements of any type. Here's the class header with the irrelevant bits stripped:
template
class Vec3
{
private:
T data[3] = {0};
public:
…
Groctel
- 130
- 11