Vincent Lefevre wrote in <20250527155503.gf128...@qaa.vinc17.org>: |On 2025-05-26 19:39:03 +0800, Jackson Helie G wrote: |> In this case (long double xx=0xf.000000000000001p0L;) tcc will lose the |> last 3 digits. If p0 is changed to p-2000, all bits will be 0. However, \ |> gcc |> and clang do not have this error. Note that my current hardware system \ |> is a |> 16-byte long double type, with 8 bits per byte. | |Note that there exist several 16-byte FP formats in practice: |x87 extended precision, double-double (a.k.a. IBM long double), |and binary128; tcc is not concerned by the second one, though. |But 0xf.000000000000001p0L is exactly representable in these |3 formats.
Not a floating-point here, so do not ask me no questions, but x87 was 10 bytes (from book reading). --steffen | |Der Kragenbaer, The moon bear, |der holt sich munter he cheerfully and one by one |einen nach dem anderen runter wa.ks himself off |(By Robert Gernhardt) _______________________________________________ Tinycc-devel mailing list Tinycc-devel@nongnu.org https://lists.nongnu.org/mailman/listinfo/tinycc-devel
