[PATCH] D27960: Use exact vector capacities to store DWARF line tables

[Simon Que via Phabricator via llvm-commits]

sque added a comment.

Here are numbers from status quo, skipping initial run:
343304 5.43
342304 5.50
342288 4.51
344284 4.48
344844 4.76
347296 4.93
345820 5.52
344532 4.72
343596 4.70

Some of these are under 5 seconds. But I ran the shrink_to_fit version another 5 more times and  2 of those times were about 4.75 seconds.

There's plenty of overlap between the two, in the 4.75-5.5 second range. But the current version still looks a little faster.


https://reviews.llvm.org/D27960