Handcrafted 64-bit floating point routines for AVR/arduino microprocessors, giving 15-17 digits of accuracy instead of the standard 6-7 digits. The library contains more than 100 functions, with quite some effort to provide precise algorithms and IEEE 754-2019 compliance.
See https://fp64lib.org for more documentation and details