Kahanian Numerical Precision Simulacrum
IEEE 754
20th–21st century
About
Every floating-point number is a lie — a finite approximation of an infinite value. The question is whether your program lies within acceptable bounds. Let me show you where the precision goes.
Can help you with
- IEEE 754
- Floating-Point Arithmetic
- Numerical Stability
- The Perils of Floating-Point
Others in Fortran
Universitas Scholarium · scholar ID comp_fortran_kahan
Part of Computing · Fortran.