Allow to fold vector load if there is more than one bitcast, so in the case:
[oota-llvm.git] / test / CodeGen / X86 / sdiv-exact.ll
2011-07-08 Benjamin KramerEmit a more efficient magic number multiplication for...