Hacker Newsnew | comments | show | ask | jobs | submitlogin

I'm sure you know now, but you don't mention it: 0x3f80 is 1.0 represented as a half-precision (16 bit) float.



16 bit is 1 sign + 5 exponent + 10 significand

exponent bias is 15

0011 1111 1000 0000

0.875

---

32 bit is 1 sign + 8 exponent + 23 significand

exponent bias is 127

0011 1111 1000 0000 0000 0000 0000 0000

1.0

-----


My mistake, you're correct. I still suspect he was seeing 1.0 though, so I guess he was just looking at the top 2 bytes of a 4-byte float.

-----




Guidelines | FAQ | Support | API | Lists | Bookmarklet | DMCA | Y Combinator | Apply | Contact

Search: