[LLVMdev] On LLD performance

Rafael EspĂ­ndola rafael.espindola at gmail.com
Sat Mar 21 08:07:23 PDT 2015


> I still think we need at least two passes if a symbol is resolved using a
> member file (first iteration resolves the symbol, and the second iteration
> is to make sure that no more symbols can be resolved using the archive
> file.) Two passes is many compared to hash table lookup, no?
>
> It may worth to mention that the hash table construction is paralellized. We
> create a hash table for each archive file in parallel because there's no
> dependency between them.

I still seems more work than is needed. It is pretty common to have
most .o in a .a used, so so the set of symbols one has to look at
quickly diminishes.

Cheers,
Rafael



More information about the llvm-dev mailing list