fp64

Code cleanup suggestions

1

### Platform Linux / BSD ### Compiler and build tool versions GCC 10.2.1 ### Operating system version GNU/Linux ### Build commands used ./b.sh ### What happens Minor stuff, feel free...

See #1391. Changelist: + Add `stb_perlin_noise3_derivatives` (return value matches `stb_perlin_noise3_wrap_nonpow2`, derivatives go into `float*` args, if not `NULL`). + Make `stb_perlin_noise3_wrap_nonpow2` faster if all 3 `wrap_*` are powers of 2...

Now that `stb_perlin.h` is back (yes!) it would be nice to have analytical derivatives of Perlin noise available available out-of-the-box (I assume a bunch of people would appreciate it, not...

Consider adding another inversible function

21

The "xorsquare" function (which I briefly describe [here](https://github.com/Marc-B-Reynolds/Marc-B-Reynolds.github.io/issues/13#issuecomment-1246966960)) is of the form: ```c++ x=((x+a)&-2)^(x*x); ``` for any integer `a`. Note: alternative version (`x^=((x+a)*(x+a))&-2;`) produces identical results for values of `a`...

Quadratic(...) fails to solve 1*x^2+0*x+0=0

1

Perhaps "fails" is a bit harsh. As a simpe [test](https://godbolt.org/z/91Kjhfcoj) shows, it finds the root (styled as `-0`), but claims that the other root is NaN. This is, of course,...

The [previous topic](https://github.com/hrydgard/ppsspp/issues/16946) ([how it started](https://github.com/hrydgard/ppsspp/pull/16817#issuecomment-1403669829)) is old, so let's consider it archived, and restart the discussion here. The story so far: 1. The transcendental FPU instructions from the previous...