Implement the rest of the shift intrinsic functions

[marktwtn]

The following shift intrinsic functions are not implemented yet.

• _mm_sll_epi*
• _mm_srl_epi*
• _mm_sra_epi*
• _mm_sllv_epi*
• _mm_srlv_epi*
• _mm_srav_epi*

However, the related shift intrinsic functions of x86 and ARM have different behaviour when the shift value is too large .
Please check the discussion in #17 (comment).

[marktwtn]

The _mm_sllv_epi*, _mm_srlv_epi* and _mm_srav_epi* are not included in SSE.
Hence, we should ignore them.

After I implement the lacking tests of _mm_sll_epi*, _mm_srl_epi* and _mm_sra_epi*, the issue should be closed.