Changes between Initial Version and Version 1 of Ticket #12506, comment 10


Ignore:
Timestamp:
Oct 26, 2016 10:35:35 AM (3 years ago)
Author:
simonpj
Comment:

Legend:

Unmodified
Added
Removed
Modified
  • Ticket #12506, comment 10

    initial v1  
    1 I clarify that I suspect the large term counts here are really the result of the library itself making heavy use of type classes, essentially making this yet another manfestation of #8095. The high desugaring times observed in compiling tree are due to the fact that we run `simpleOptPgm` during desugaring.
     1I clarify that I suspect the large term counts here are really the result of the library itself making heavy use of type functions, essentially making this yet another manfestation of #8095. The high desugaring times observed in compiling tree are due to the fact that we run `simpleOptPgm` during desugaring.
    22
    33I didn't have a chance to look at compilation time regressed from 7.8 to 7.10 to 8.0. When I get a chance the best next step will likely to focus on the 7.10 to 8.0 regression (since it is both newer (and therefore easier to debug) as well as larger). First verify that the desugared Core is equivalent, then look at compiler profiles to see if the compiler implementation has regressed.