cmd/compile,math: improve performance for some math functions for ppc64x

[laboger]

Please answer these questions before submitting your issue. Thanks!

What version of Go are you using (go version)?

go tip

What operating system and processor architecture are you using (go env)?

Ubuntu (or any) ppc64le

Making some improvements to several math functions by writing some in asm and implementing some as intrinsics.

I will be creating multiple CLs to do this, so instead of opening multiple issues, I would like to assign them all to this one.

The first change will be to implement the trunc, floor, and ceil functions as intrinsics.

[ALTree]

This is (a subset of) #8913 (math, crypto, hash: Write power64 versions of assembler routines where applicable)

[gopherbot]

Change https://golang.org/cl/54654 mentions this issue: cmd/compile: intrinsics for trunc, floor, ceil on ppc64x

[gopherbot]

Change https://golang.org/cl/59932 mentions this issue: cmd/compile, math/bits: intrinsics for rotates on ppc64le

[gopherbot]

Change https://golang.org/cl/63290 mentions this issue: cmd/compile,math: improve int<->float conversions on ppc64x

[gopherbot]

Change https://golang.org/cl/67130 mentions this issue: cmd/compile,math,cmd/internal/obj/ppc64: abs,copysign instrinsics on ppc64x

[gopherbot]

Change https://golang.org/cl/74411 mentions this issue: math: improve dim, min, max, modf for ppc64x

[laboger]

I have merged all the math changes I had planned for go 1.10, so will close this one.