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

Also very common in embedded systems that simply do not have any floating point capabilities at all. For my CS dissertation I had to implement part of an OpenGL pipeline on a VLIW DSP designed for video coding, and all it had was 32-bit integer ALU’s. Dealing with dynamic range was pretty much half of the total effort to map the reference code that used floating point to fixed point with mixed precision.


Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: