Closed russell-taylor closed 3 years ago
Doing all of the floating-point number parsing was slowing things down. Made the new code cache the lines and parse them only as needed. This reduced the loading time by about a second.
Modified the code to only load and parse the HET database if it ever needs to do a lookup. Also modified the code to not reduce the list if it is not either adding hydrogens or optimizing. This makes the new code faster than the old when running with -TRIM.
Fixed in version 4.2.
It seems to be much slower, perhaps since we added the shared_ptr approaches (like 2 vs. 5 seconds on 1A0F.pdb).