This is the mail archive of the gcc-patches@gcc.gnu.org mailing list for the GCC project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

[tree-ssa] Speeding up CCP


This small change gets us from 701 seconds to 691 seconds of user
time on my collection of testcases.   FWIW, this brings us back
to the compile-time performance of the branch before DCE was 
enabled by default :-)

	* tree-ssa-ccp.c (tree_ssa_ccp): Work through the entire
	ssa_edges worklist each iteration through the main loop.

Index: tree-ssa-ccp.c
===================================================================
RCS file: /cvs/gcc/gcc/gcc/Attic/tree-ssa-ccp.c,v
retrieving revision 1.1.2.47
diff -c -3 -p -r1.1.2.47 tree-ssa-ccp.c
*** tree-ssa-ccp.c	14 Feb 2003 05:34:58 -0000	1.1.2.47
--- tree-ssa-ccp.c	14 Feb 2003 16:09:37 -0000
*************** tree_ssa_ccp (fndecl)
*** 155,161 ****
  	  simulate_block (dest_block);
  	}
  
!       if (VARRAY_ACTIVE_SIZE (ssa_edges) > 0)
  	{
  	  /* Pull the next reference off the worklist.  The SSA edges
  	     worklist stores the origination definition for each edge.  */
--- 155,163 ----
  	  simulate_block (dest_block);
  	}
  
!       /* The SSA_EDGES worklist can get rather large.  Go ahead and
!          drain the entire worklist each iteration through this loop.  */
!       while (VARRAY_ACTIVE_SIZE (ssa_edges) > 0)
  	{
  	  /* Pull the next reference off the worklist.  The SSA edges
  	     worklist stores the origination definition for each edge.  */







Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]