Extend x86 inline asm to support ymm and zmm vector registers

[CathalMullan]

Existing code only handled xmm registers, this adds support for ymm and zmm.

Resolves #1616

[bjorn3]

Could you add a test exercising this new code to example/std_example.rs behind #[cfg(all(target_arch = "x86_64, not(jit)))]?

[bjorn3]

Thanks!