Jump to content
  • Advertisement
Sign in to follow this  
coderchris

Converting to/from 10/11 bit floating point number

This topic is 3537 days old which is more than the 365 day threshold we allow for new replies. Please post a new topic.

If you intended to correct an error in the post then please contact us.

Recommended Posts

Hello, Im curious if anyone could explain how to go to/from a 10/11 bit float to/from a standard 32 bit float Both 10 and 11 have no sign bit, and 5 bits of exponent. 10 has then 5 bits of mantissa and 11 has 6 bits of mantissa. Obviously there is no sign bit, so I just need to drop the sign off of the 32 bit float. I believe I also understand how to convert the exponent. bias for a 8 bits of exponent = 127 bias for 5 bits of exponent = 15 So the formula to convert from exp32 to exp10/11 is then exp10/11 = (exp32 - 127) + 15 Heres where im not quite sure what to do. if exp10/11 <= 0, I know something special happens to convert it to denormalized or something. If exp10/11 >= 31 then its too large to store, so i guess we set the new float to infinity? And what do we do to the mantissa even when 0 < exp10/11 < 31? Thanks, Chris

Share this post


Link to post
Share on other sites
Advertisement
Sign in to follow this  

  • Advertisement
×

Important Information

By using GameDev.net, you agree to our community Guidelines, Terms of Use, and Privacy Policy.

We are the game development community.

Whether you are an indie, hobbyist, AAA developer, or just trying to learn, GameDev.net is the place for you to learn, share, and connect with the games industry. Learn more About Us or sign up!

Sign me up!