Specifying Math Libraries

optimized-routines is an optimized implementation for library functions of the Arm processor. BiSheng compiler integrates optimized-routines into the toolchain as a dynamic link library. optimized-routines supports both scalar and vector implementation of mathematical functions.

Run the following command to specify a scalar math library:

1

clang -O3 -lmathlib -lm

Run the following command to specify a vector math library:

1

clang -O3 -fveclib=MATHLIB -lmathlib -lm

Some mathlib functions are not supported. The -lm option needs to be added to link the libm. The mathlib must be linked before the libm.