based on IEEE floating-pointing format.
Format A:
- 1 sign bit.
- 5 epx bits. bias = 15.
- 3 fraction bits.
| Bits | Value |
|---|---|
| 1 01111 001 | -9/8 |
| 0 10110 011 | 176 |
| 1 00111 010 | -5/1024 |
| 0 00000 111 | 7/2^{14} |
| 1 11100 000 | -2^{13} |
| 0 10111 100 | 384 |
Format B:
- 1 sign bit.
- 4 epx bits. bias = 7.
- 4 fraction bits.
| Bits | Value |
|---|---|
| 1 0111 0010 | -9/8 |
| 0 1110 0110 | 176 |
| 1 0000 0101 | -5/1024 |
| 0 0000 0001 | 1/2^{10} |
| 1 1111 0000 | -inf |
| 0 1111 0000 | inf |