Bug in value-prof.c:visit_hist

[Robert Kidd]

There appears to be a bug in value-prof.c:visit_hist rev 121554.   
This function always returns 0, which causes htab_traverse to exit  
early.  This means that only the first histogram that appears in cfun- 
>value_histograms->entries is ever checked, so verify_histograms  
will only indicate an error if the first histogram is unreachable.   
The attached patch changes the return value to ensure that all  
histograms are checked.  This patch bootstraps and passes make check  
on x86_64.


Robert Kidd
[EMAIL PROTECTED]

Index: gcc/value-prof.c
===
--- gcc/value-prof.c(revision 121671)
+++ gcc/value-prof.c(working copy)
@@ -353,8 +353,9 @@ visit_hist (void **slot, void *data)
   dump_histogram_value (stderr, hist);
   debug_generic_stmt (hist->hvalue.stmt);
   error_found = true;
+  return 0;
 }
-  return 0;
+  return 1;
 }
 
 /* Verify sanity of the histograms.  */

GCC superblock and region formation support

[Robert Kidd]

As a quick introduction, my name is Robert Kidd, and I'm working with 
the Gelato Federation to improve the performance of GCC on Itanium.  In 
particular, I'm looking into improving GCC's superblock support, 
hopefully bringing over some of what we have learned with the IMPACT 
compiler project.

After studying GCC's current tail duplication and extended basic block 
scheduling code, I'm thinking about ways to improve GCC's 
representation.  I'm aiming for a more concrete form than a basic block 
trace.  Steven Bosscher pointed me in the direction of the region 
formation project by Daniel Berlin and Kenneth Zadeck, which sounds 
like a good basis for a superblock representation.  What is the status 
of this project?  Has any documentation or code been released?

Thanks
Robert Kidd
[EMAIL PROTECTED]

[ia64-improvements] Moved Superblock formation to Tree-SSA

[Robert Kidd]

I checked a patch into the ia64-improvements branch to move  
Superblock formation to the Tree-SSA level.  This patch also brings  
an update to the trace formation pass.  When building a trace inside  
a loop, we now follow the backedge once to add the loop header to the  
trace.  This allows tail duplication to form a superblock loop that  
can be processed by the loop unroller later on.


2006-02-17  Robert Kidd  <[EMAIL PROTECTED]>

* tracer.c (mark_bb_seen): New function.
(bb_seen_p): New function.
(count_insns): Use estimate_num_insns instead of counting RTL
instructions.
(find_trace): Tracing a loop now traverses the back edge once
to duplicate the loop header.
(tail_duplicate): Phi nodes are fixed after duplicating
blocks.
(layout_superblocks): Remove function.
(execute_tracer): CFG layout stuff is not needed at the tree
level.
(rest_of_handle_tracer): Remove function.
* bb-reorder.c (rest_of_handle_reorder_blocks): Removed call
to RTL level tracer.
* passes.c (init_optimization_passes): Moved pass_tracer from
after pass_rtl_ifcvt to after pass_build_ssa.

Robert Kidd
[EMAIL PROTECTED]