Hacker News new | ask | show | jobs
by nwhitehead 3262 days ago
> If you’re relying on c99 complex or std::complex and fast complex operations, and you don’t have -fcx-limited-range or -ffast-math turned on, you should probably consider this a bug in your project.

I would rephrase this as, "If you care about floating point performance and have not investigated the compiler flags related to floating point performance, consider this a bug in your project."