memo_int_sqrt
| Crates.io | memo_int_sqrt |
| lib.rs | memo_int_sqrt |
| version | 0.1.1 |
| created_at | 2019-02-04 21:28:31.783537+00 |
| updated_at | 2019-02-05 21:58:33.326594+00 |
| description | Memoized square root and inverse square root for integers |
| homepage | https://github.com/stevebob/memo-int-sqrt.git |
| repository | https://github.com/stevebob/memo-int-sqrt.git |
| max_upload_size | |
| id | 112765 |
| size | 9,204 |
Stephen Sherratt (gridbugs)
documentation
https://docs.rs/memo_int_sqrt
README
memo_int_sqrt
Definitely run the benchmarks before using this. In my benchmarking, I found that for f32s, .sqrt().recip() is often
faster than using the lookup table. For f64s, the lookup table is faster for inverse squareroot, but not for squareroot.