Dataflow analysis rework

Rework key portions of the dataflow analysis code to avoid the
massive slowdowns we're seeing with methods containing more than
20,000 basic blocks.

Add a somewhat hacky lookup cache in front of findBlock.  Later
we'll want to rework the "GrowableList" mechanism to incorporate
this properly.

Add memory to the bit vector utilities to speed up the iterators
for large sparse vectors. (Similarly, in the future we'll want
to include a rewrite of the home-grown bit-vector manipluation
utilites to be better with large ones).

With this CL, compilation speed on the degenerate cases is > 10x
faster.  Compilation of typically-sized methods will see a smallish
improvement.

Change-Id: I7f086c1229dd4fe62f0a5fc361234bf204ebc2b1

1 files changed, 12 insertions, 0 deletions

diff --git a/src/compiler/CompilerUtility.h b/src/compiler/CompilerUtility.h
index d7612ffc2f..a68ee5fb0a 100644
--- a/src/compiler/CompilerUtility.h
+++ b/src/compiler/CompilerUtility.h
@@ -58,8 +58,20 @@ typedef struct GrowableListIterator {
  */
 struct ArenaBitVector {
     bool    expandable;     /* expand bitmap if we run out? */
+    bool    firstDirty;     /* when true, don't believe firstBitSet */
+    bool    lastDirty;      /* when true, don't believe lastBitSet */
     u4      storageSize;    /* current size, in 32-bit words */
     u4*     storage;
+                            /*
+                             * Opportunistically remember first and
+                             * last set bits.  This yeilds a performance
+                             * advantage in cases where large
+                             * sparse vectors are repeatedly scanned
+                             * (something that can happen a lot during
+                             * dataflow analysis.
+                             */
+    int     firstBitSet;
+    int     lastBitSet;
 };
 
 /* Handy iterator to walk through the bit positions set to 1 */