[libc][math] Added atanf function.

[orex]

Summary

Performance by core-math (core-math/glibc 2.31/current llvm-14):
28.879/20.843/20.15

The revision should be applied on top of D132811.

[zimmermann6]

ok for me, I get slightly different figures on a AMD EPYC 7282:

zimmerma@biscotte:~/svn/core-math$ LIBM=/localdisk/zimmerma/llvm-project/build/projects/libc/lib/libllvmlibc.a CORE_MATH_PERF_MODE=rdtsc ./perf.sh atanf
GNU libc version: 2.34
GNU libc release: stable
17.539
31.797
26.903

[lntue]

libc/src/__support/FPUtil/FPBits.h

200

Please add the same function for x86_64/LongDoubleBits

libc/src/math/generic/atanf.cpp

25

Are these constants affected by rounding modes? Also it would be better to define this constant ourselves in some our own header, not relying on the constants defined in math.h.

30

Add comment about its hexadecimal float value.

32

Add comment about its hexadecimal float value.

38

Add comment about its hexadecimal float value.

42

Add comment about its hexadecimal float value.

45

Add comment about its hexadecimal float value.

[orex]

Review fixes.

[lntue]

libc/src/math/generic/atanf.cpp

13

Update this header file.