math: optimize Exp and Exp2 on arm64

math: optimize Exp and Exp2 on arm64

This CL implements Exp and Exp2 with arm64 assembly. By inlining Ldexp and
using fused instructions(fmadd, fmsub, fnmsub), this CL helps to improve
the performance of functions Exp, Exp2, Sinh, Cosh and Tanh.

Benchmarks:
name                   old time/op  new time/op  delta
Cosh-8                  138ns ± 0%    96ns ± 0%  -30.72%  (p=0.008 n=5+5)
Exp-8                   105ns ± 0%    58ns ± 0%  -45.24%  (p=0.000 n=5+4)
Exp2-8                  100ns ± 0%    57ns ± 0%  -43.21%  (p=0.008 n=5+5)
Sinh-8                  139ns ± 0%   102ns ± 0%  -26.62%  (p=0.008 n=5+5)
Tanh-8                  134ns ± 0%   100ns ± 0%  -25.67%  (p=0.008 n=5+5)

Change-Id: I7483a3333062a1d3525cedf3de56db78d79031c6
Reviewed-on: https://go-review.googlesource.com/86615
Run-TryBot: Cherry Zhang <cherryyz@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Cherry Zhang <cherryyz@google.com>