Beating 64-bit IEEE floating points using a 32-bit data format

by ityonemo
GNU/Linux ◆ xterm ◆ bash 1470 views

Using a “quire”, which is a datastructure intended to be implemented on binary hardware, I show how you can do exact solutions of systems of linear equations.

SigmoidNumber julia package:

https://github.com/interplanetary-robot/SigmoidNumbers

Beyond Floating Point video:

https://www.youtube.com/watch?v=aP0Y1uAA-2Y

More by ityonemo

zigler's new errors 00:37

by ityonemo

See all