Add experimental precision doubling transform (#3465)
This PR adds an experimental precision doubling transform, following the basic approach outlined in Dekker 1971 ([pdf](http://csclub.uwaterloo.ca/~pbarfuss/dekker1971.pdf)). When this transform is applied, the number of significant bits is approximately doubled compared to the base operation.
Simple demo:
```python
In [1]: import jax.numpy as jnp
In [2]: from jax.experimental.doubledouble import doubledouble
In [3]: def f(a, b):
...: return a + b - a
...:
In [4]: f(1E20, 1.0) # float64 loses precision
Out[4]: 0.0
In [5]: g = doubledouble(f)(1E20, 1.0)
Out[5]: DeviceArray(1., dtype=float64)
```
This initial experiment supports basic arithmetic operators and inequalities.
