Compiler-Assisted Floating-Point Error Analysis and Profiling with FPChecker

I was expecting them to be aware of https://herbie.uwplse.org/ Herbie but it seems like either they don't know of it or they don't see it as relevant.

My hope for the future is that Herbie's approach can evolve to a point where the person writing the software doesn't directly write FPU operations, any more than you'd manually work out all the best ways to optimise integer arithmetic. You write what you meant and the compiler figures out how to deliver that with FPU instructions. The parameters are more complicated for the second operation, but it's surely tractable.

"I want to calculate function F(x) here, I care about 0 <= x <= 4.0, inside those limits we need +/- 0.1% accuracy, for other values I don't care, and I want you to go as fast as possible within those parameters".

FPChecker looks very cool. I wish the exponent graph had the option to show base 2 though, so it's slightly easier to map it back to the underlying float representation.

Is there actually video of the talk?