Skip to main content

Floating Point

This article is a work in progress.

Precision​

About Comparisons​

Due to small rounding errors, exact equality is often not a reliable way to compare floating point numbers. Therefore the default equals operation checks that two numbers are within a small delta.

There is also an exactEquals method that compares the floating point values directly.

return Math.abs(a - b) <= config.EPSILON * Math.max(1.0, Math.abs(a), Math.abs(b));
  • Floating point comparison implementation in gl-matrix and alternatives (links)
  • Comparisons