jm*** 发表于 2024-1-18 04:35
This info may help those wanting 64bit (8 byte) support, SDCC added that in 2018.
February 26, 201 ...
Early single precision floating-point numbers in C51 were non reentrant (using simulation stacks), but later optimized to be reentrant (all other transcendental functions were non reentrant).
There are too few available registers in C51, and it is almost impossible to develop a concise reentrant double precision floating-point function library.