Check libm vs Eps1Over1e3Erf performance

[hombit]

Currently it looks like they have the same performance and we can safely drop Eps1Over1e3Erf.

cargo bench

Benchmarking erf_eps_1over1e3 for f32: Collecting 100 samples in estimated 5.0000 s (2
erf_eps_1over1e3 for f32
                        time:   [1.9824 ns 1.9927 ns 2.0038 ns]
Found 7 outliers among 100 measurements (7.00%)
  3 (3.00%) high mild
  4 (4.00%) high severe

Benchmarking libm::erf(f) for f32: Collecting 100 samples in estimated 5.0000 s (2.5B
libm::erf(f) for f32    time:   [1.9677 ns 1.9789 ns 1.9913 ns]
Found 3 outliers among 100 measurements (3.00%)
  3 (3.00%) high mild

But also

Benchmarking DmDt<f32>::convert_lc_to_gausses, erf type is Eps1Over1e3Erf: Warming up
Benchmarking DmDt<f32>::convert_lc_to_gausses, erf type is Eps1Over1e3Erf: Collecting
DmDt<f32>::convert_lc_to_gausses, erf type is Eps1Over1e3Erf
                        time:   [184.31 µs 184.62 µs 184.96 µs]
Found 3 outliers among 100 measurements (3.00%)
  1 (1.00%) high mild
  2 (2.00%) high severe

Benchmarking DmDt<f64>::convert_lc_to_gausses, erf type is ExactErf: Warming up for 3.0000 s
Warning: Unable to complete 100 samples in 5.0s. You may wish to increase target time to 5.1s, enable flat sampling, or reduce sample count to 60.
Benchmarking DmDt<f64>::convert_lc_to_gausses, erf type is ExactErf: Collecting 100 sa
DmDt<f64>::convert_lc_to_gausses, erf type is ExactErf
                        time:   [1.0261 ms 1.0298 ms 1.0337 ms]
Found 3 outliers among 100 measurements (3.00%)
  1 (1.00%) low mild
  2 (2.00%) high mild