[Issue 10866] New: Regression (2.064 git-head) Massive compiler slowdown

[d-bugmail at puremagic.com]

http://d.puremagic.com/issues/show_bug.cgi?id=10866

           Summary: Regression (2.064 git-head) Massive compiler slowdown
           Product: D
           Version: D2
          Platform: All
        OS/Version: All
            Status: NEW
          Keywords: performance
          Severity: regression
          Priority: P2
         Component: DMD
        AssignedTo: nobody at puremagic.com
        ReportedBy: andrej.mitrovich at gmail.com


--- Comment #0 from Andrej Mitrovic <andrej.mitrovich at gmail.com> 2013-08-21 12:47:19 PDT ---
What's up with the slowdown of the compiler in git-head? Here is the
average time to build this simple test-file:

-----
module test;
import std.stdio;
void main() { }
-----

2.063.2:
392 msecs

Git-head (1a6da16ef112e03607574758b722698f2950b0de):
1_251 msecs

The timings are pretty stable from my end.

In another one of my projects the timings are:
3_500 ms in 2.063
4_800 ms in 2.064

The codebase is around 10K.

However in other projects I get a speedup. Here's WindowsAPI being built:

2.063: 19_903 msecs.
2.064: 9_782 msecs.

I guess Walter's recent speedups made the compiler faster. 

But there must have been acommit which seems to have slowed some other things
down. According to Don it may have been this pull:
https://github.com/D-Programming-Language/dmd/pull/2421#issuecomment-22688675

And I can confirm it from my side. In the 10K project I've tried running a full
build with these two DMD commits:

commit A: 805dd761dd85c98d60b3957bd5679cdbbf23c7f2
commit B: 52cc28752a6fd01446fd09a7ab9bd9f94ea601d2

commit A is the one introduced by Pull request #2421, commit B is one commit
before it.

Measurements:

commit A: 4_795 msecs
commit B: 3_900 msecs

-- 
Configure issuemail: http://d.puremagic.com/issues/userprefs.cgi?tab=email
------- You are receiving this mail because: -------