[X86] Improved lowering of packed vector shifts to vpsllq/vpsrlq.
[oota-llvm.git] / test / CodeGen / X86 / imul64-lea.ll
2013-02-01 David SehrTwo changes relevant to LEA and x32: