dear: now, i see that problem for half-precision floating-point value. for intel  IA-32, there are not its instructin.                      niu wensheng                       Mar, 3, 2023 ---Original--- From: "Paul Koning via Gdb"