Floating-point instructions affect the following bits in the status word: CC 1 and
CC 0, OV and OS.
The following tables show the signal state of the bits in the status word for the results of instructions with floating-point numbers (32 bits):
Valid Range for the Result | CC 1 | CC 0 | OV | OS |
+0, -0 (zero) | 0 | 0 | 0 | * |
-3.402823E+38 < result < -1.175494E-38 (negative number) | 0 | 1 | 0 | * |
+1.175494E-38 < result < 3.402824E+38 (positive number) | 1 | 0 | 0 | * |
* The OS bit is not affected by the result of the instruction.
Invalid Range for the Result | CC 1 | CC 0 | OV | OS |
Underflow -1.175494E-38 < result < - 1.401298E-45 (negative number) | 0 | 0 | 1 | 1 |
Underflow +1.401298E-45 < result < +1.175494E-38 (positive number) | 0 | 0 | 1 | 1 |
Overflow Result < -3.402823E+38 (negative number) | 0 | 1 | 1 | 1 |
Overflow Result > 3.402823E+38 (positive number) | 1 | 0 | 1 | 1 |
Not a valid floating-point number or illegal instruction (input value outside the valid range) | 1 | 1 | 1 | 1 |
0 comments:
Post a Comment