[NVPTX] Add intrinsics for shfl instructions.

Summary:
Currently clang emits these instructions via inline (volatile) asm in
the CUDA headers.  Switching to intrinsics will let the optimizer reason
across calls to these intrinsics.

Reviewers: tra

Subscribers: llvm-commits, jholewinski

Differential Revision: http://reviews.llvm.org/D21160

llvm-svn: 272298