SPRUJB6B November 2024 – May 2025 AM2612
The treatment of the various IEEE floating-point numerical formats for this TMU are the same as the FPU implementation.
The encoding of the floating-point formats is given in the below table
| S32 | E32 (7:0) | M32 (22:0) | Value (V) |
|---|---|---|---|
| 0 | 0 | 0 | Zero (V = 0) |
| 1 | 0 | 0 | Negative Zero (V = -0) |
|
0 +ive 1 -ive |
0 | Non-zero | De-normalized (V=(-1)S* 2(-126)* (0.M)) |
|
0 +ive 1 -ive |
1 to 254 | 0 to 0x7FFFFF | Normal Range (V=(-1)S * 2(E-127) * (1.M)) |
| 0 | 254 | 0x7FFFFF | Positive Max (V = +Max) |
| 1 | 254 | 0x7FFFFF | Negative Max (V = -Max) |
| 0 | Maximum = 255 | 0 | Positive Infinity (V = +Infinity) |
| 1 | Maximum = 255 | 0 | Negative Infinity (V = -Infinity) |
| X | Maximum = 255 | Non-zero | Not A Number (V = NaN) |