Re: [PPL-devel] Floating-point rounding mode