[PATCH] [X86][SSE] Vector integer/float conversion memory folding (cvttps2dq / cvttpd2dq)

[Simon Pilgrim]

Many of the Int_CVT and CVT 'scalar' folds look suspicious and I wish to more thoroughly test them, but would prefer to do that separately from this patch.

I found some rather poor code generation for non-AVX code for double -> int32 that needs fixing as well.

http://reviews.llvm.org/D6001