Use CRT float functions from MathF

[tijoytom-zz]

Right now MathF functions are just forwards to Math , but we need to change them to use lower precision but efficient f* functions..

[tannergooding]

@jkotas, the changes here look relatively simple to make. But I'm not familiar with the RuntimeLibrary definition (

corert/src/System.Private.CoreLib/src/System/Runtime/RuntimeImports.cs

Line 796 in 143634a

[RuntimeImport(RuntimeLibrary, "asin")]

), where would I find it?

[jkotas]

These match functions has been coming straight from CRT. There is nothing defining them explicitly.

[tannergooding]

Awesome. Then this looks like a relatively simple change. I'm willing to take this on if someone else isn't going to 😄