Opened 11 years ago

Closed 11 years ago

Last modified 10 years ago

#2720 closed bug (fixed)

eyeball/inline1 still isn't optimised with -fno-method-sharing

Reported by: rl Owned by:
Priority: normal Milestone: 6.10 branch
Component: Compiler Version: 6.10.1
Keywords: Cc:
Operating System: Unknown/Multiple Architecture: Unknown/Multiple
Type of failure: Runtime performance bug Test Case:
Blocked By: Blocking:
Related Tickets: Differential Rev(s):
Wiki Page:

Description

Just as a reminder: eyeball/inline1.hs isn't optimised properly if -fno-method-sharing is set after this patch:

Tue Sep  9 08:50:11 PDT 2008  simonpj@microsoft.com
 * Important performance wibble to callSiteInline (the n_vals_wanted > 0 thing)

This affects DPH code quite a bit.

Change History (3)

comment:1 Changed 11 years ago by igloo

difficulty: Unknown
Milestone: 6.10 branch

comment:2 Changed 11 years ago by simonpj

Resolution: fixed
Status: newclosed

I believe I've fixed this

Tue Oct 28 14:08:28 GMT Standard Time 2008  simonpj@microsoft.com
  * Fix Trac #2720: inlining and casts
  
  The issue here is what happens when we have
  
  	(f |> co) x
  
  where f is itself marked INLINE.  We want callSiteInline to "see" 
  the fact that the function is applied, and hence have some incentive
  to inline.  I've done this by extending CoreUnfold.CallCtxt with 
  ValAppCtxt.  I think that should catch this case without messing up
  any of the others.

Do try it out. I'll close it meanwhile.

Simon

comment:3 Changed 10 years ago by simonmar

Type of failure: Runtime performance bug
Note: See TracTickets for help on using tickets.