fix some shift operations to agree with Base (and prevent undefined behavior)

[KristofferC]

…ehavior)

Due to an off by one, shifting with the exact amount as the number of bits caused garbage to get returned:

julia> f2(v) = v >> 64
f2 (generic function with 1 method)

julia> v = Vec(1,2,3,4)
<4 x Int64>[1, 2, 3, 4]

julia> @code_llvm f2(v)
;  @ REPL[25]:1 within `f2`
define void @julia_f2_542([1 x <4 x i64>]* noalias nocapture noundef nonnull sret([1 x <4 x i64>]) align 32 dereferenceable(32) %0, [1 x <4 x i64>]* nocapture noundef nonnull readonly align 16 dereferenceable(32) %1) #0 {
top:
  ret void
}

Ref https://discourse.julialang.org/t/simd-struggles-seeking-solutions-with-kangarootwelve-jl/105800/23?u=kristoffer.carlsson

[codecov-commenter]

Codecov Report

All modified and coverable lines are covered by tests ✅

❗ Your organization needs to install the Codecov GitHub app to enable full functionality.

Files	Coverage Δ
src/simdvec.jl	79.12% <100.00%> (ø)

... and 1 file with indirect coverage changes

📢 Thoughts on this report? Let us know!

[KristofferC]

Error on nightly is unrelated.