Searched refs:x_is_nan (Results 1 – 1 of 1) sorted by relevance
161 llvm::Value* x_is_nan = b->CreateFCmpUNO(x, x); in EmitReducePrecisionFloat() local164 result = b->CreateSelect(x_is_nan, x, result); in EmitReducePrecisionFloat()167 x_is_nan, llvm::ConstantFP::getInfinity(float_type), result); in EmitReducePrecisionFloat()