optimize bitcast(trunc(bitcast(x))) where the result is a float and 'x'
[oota-llvm.git] / test / Transforms / InstCombine / fcmp-special.ll
2010-02-22 Dan GohmanConstant-fold certain comparisons with infinity and...