| commit | 1bdd5c8b9868ebc092074e078604acb80546e43f | [log] [tgz] |
|---|---|---|
| author | Rich Felker <dalias@aerifal.cx> | Wed Apr 11 14:11:47 2012 -0400 |
| committer | Rich Felker <dalias@aerifal.cx> | Wed Apr 11 14:11:47 2012 -0400 |
| tree | a9d2120c543db826cc9c30c902b62bd23267c10a | |
| parent | 7ef1a9bba56aa756d8166c4c93cf4a178d6c0c0c [diff] |
optimize floatscan: avoid excessive upscaling upscaling by even one step too much creates 3-29 extra iterations for the next loop. this is still suboptimal since it always goes by 2^29 rather than using a smaller upscale factor when nearing the target, but performance on common, small-magnitude, few-digit values has already more than doubled with this change. more optimizations on the way...