This patch adds functionality '#pragma clang loop vectorize_assume_alignment'. The load/store access in for loop following the '#pragma clang loop vectorize_assume_alignment(n)' uses n bit for align access.
#pragma clang loop vectorize_assume_alignment(n) for(){ .... }
This pragma depends on D69900
What does this actually mean, in the end?
Will it assume that whatever alignment is required
for whatever vectorization width chosen,
is the actual alignment?