fast-math

fast-math copied to clipboard

There are hardware instructions on X86 for approximate inverse square root and square root which can improved upon via newton's method. See the for loop at the bottom of `bodies_advance`...

tkaitchuck

Implement fast tanh

3

Partially adresses #1.

vks

Match `std`'s API

1

All of the interesting functions in https://doc.rust-lang.org/std/primitive.f32.html: - [ ] #8 trig - [ ] #9 inverse trig - [ ] #10 exp_m1 - [ ] #11 log_1p - [...

huonw

- [ ] `log` (with `base` argument) - [ ] `ln` - [ ] `log10` Also their `_1p` variants (#11).

huonw

Potentially best implemented in the form of `log2_1p`, but `ln_1p` should be exposed too.

huonw

Potentially also `exp2_m1` etc, since packaging that up likely saves a multiplication?

huonw

- [ ] `erf` - [ ] `erfc`

huonw

- [ ] `lgamma` - [ ] `tgamma` These names are the conventional/libc ones, but it might be nice to break with convention?

huonw

- [ ] `powf` - [ ] `powi`

huonw

- [ ] `asinh` - [ ] `acosh` - [ ] `atanh`

huonw