[llvm-commits] CVS: llvm/lib/CodeGen/README.txt

Dale Johannesen dalej at apple.com
Fri May 18 11:47:00 PDT 2007

Changes in directory llvm/lib/CodeGen:

README.txt updated: 1.5 -> 1.6
Log message:

Document an inefficiency in tail merging.

Diffs of the changes:  (+19 -0)

 README.txt |   19 +++++++++++++++++++
 1 files changed, 19 insertions(+)

Index: llvm/lib/CodeGen/README.txt
diff -u llvm/lib/CodeGen/README.txt:1.5 llvm/lib/CodeGen/README.txt:1.6
--- llvm/lib/CodeGen/README.txt:1.5	Mon Apr 30 13:42:09 2007
+++ llvm/lib/CodeGen/README.txt	Fri May 18 13:46:40 2007
@@ -142,3 +142,22 @@
 load [T + 7]
 load [T + 15]
+Tail merging issue:
+When we're trying to merge the tails of predecessors of a block I, and there
+are more than 2 predecessors, we don't do it optimally.  Suppose predecessors
+are A,B,C where B and C have 5 instructions in common, and A has 2 in common
+with B or C.  We want to get:
+  jmp C3
+  jmp C2
+C2:  3 common to B and C but not A
+C3:  2 common to all 3
+You get this if B and C are merged first, but currently it might randomly decide
+to merge A and B first, which results in not sharing the C2 instructions.  We 
+could look at all N*(N-1) combinations of predecessors and merge the ones with
+the most instructions in common first.  Usually that will be fast, but it 
+could get slow on big graphs (e.g. large switches tend to have blocks with many 

More information about the llvm-commits mailing list