This is a Julia interface to the IEEE 754 binary128 floating point format, backed by libquadmath (or libm for some architectures). It provides a Float128 type and associated functionality.
Quadmath supports x86_64 and AArch64.
- On AArch64 Linux and FreeBSD, libquadmath is not available, so
long doublefunctions from libm are used instead. Unlike x86_64,long doubleon AArch64 is IEEE 754 binary128. - On x86_64, it may require gcc 4.3 or later for the
__float128type. - It has not been tested on 32-bit x86 Linux.
- It does not work on 32-bit ARM due to the lack of libquadmath support for that platform.
Quadmath.jl is a registered package, so can be added via ]add Quadmath.
Thank you to Harald Hofstätter, who provided the first iteration of Quadmath.jl.