Merge pull request #16053 from e-kayrakli/static-fast-follower-minimal

Add static analysis to avoid adding slow followers in some cases

This PR adds a static analysis to detect whether zippered distributed arrays
share the same domain and thus can leverage the fast follower optimization.

This is a small part of what I attempted at #16001.

This PR also rearranges the newly-added autoLocalAccess.cpp to make it more
general purpose.

Summary of changes:

- Add `staticFastFollowerAnalysis` to `autoLocalAccess.cpp`
- Do some refactor on that file:
  - Rename to `preNormalizeOptimizations.cpp`
  - Put function declarations in the beginning of the file
  - Add a function `doPreNormalizeArrayOptimizations` to handle automatic local
    access and fast follower analysis right before normalization.

Although the main motivation for this PR was to reduce the compilation time
overheads, in some of the benchmarks I have seen that clones removed by this PR
to constitute only a small partion of those extra clones. #16001 gives an idea
of what needs to be done to remove a larger number of copies, and I hope to
revisit that effort soon.

[Reviewed by @ronawho]

Test:
- [x] add a test to lock the behavior
- [x] standard
- [x] gasnet

compiler/AST/ForallStmt.cpp

@@ -56,6 +56,7 @@ ForallStmt::ForallStmt(BlockStmt* body):
 
 
   optInfo.autoLocalAccessChecked = false;
+  optInfo.confirmedFastFollower = false;
 
   gForallStmts.add(this);
 }

compiler/AST/foralls.cpp

@@ -964,9 +964,16 @@ static void buildLeaderLoopBody(ForallStmt* pfs, Expr* iterExpr) {
                                           new_Expr("'move'(%S, %S)", T2, gFalse)));
     } else {
       leadForLoop->insertAtTail("'move'(%S, chpl__staticFastFollowCheckZip(%S))", T1, iterRec);
-      leadForLoop->insertAtTail(new CondStmt(new SymExpr(T1),
-                                          new_Expr("'move'(%S, chpl__dynamicFastFollowCheckZip(%S))", T2, iterRec),
-                                          new_Expr("'move'(%S, %S)", T2, gFalse)));
+
+      // override the dynamic check if the compiler can prove it's safe
+      if (pfs->optInfo.confirmedFastFollower) {
+        leadForLoop->insertAtTail(new_Expr("'move'(%S, %S)", T2, T1));
+      }
+      else {
+        leadForLoop->insertAtTail(new CondStmt(new SymExpr(T1),
+                                            new_Expr("'move'(%S, chpl__dynamicFastFollowCheckZip(%S))", T2, iterRec),
+                                            new_Expr("'move'(%S, %S)", T2, gFalse)));
+      }
     }
 
     SymbolMap map;

compiler/include/ForallStmt.h

@@ -46,6 +46,9 @@ class ForallOptimizationInfo {
     std::vector<Symbol *> staticCheckSymsForDynamicCandidates;
 
     bool autoLocalAccessChecked;
+
+
+    bool confirmedFastFollower;
 };
 
 ///////////////////////////////////

compiler/include/autoLocalAccess.h → compiler/include/preNormalizeOptimizations.h

@@ -18,14 +18,14 @@
  * limitations under the License.
  */
 
-#ifndef _AUTO_LOCAL_ACCESS_H_
-#define _AUTO_LOCAL_ACCESS_H_
+#ifndef _PRE_NORMALIZE_OPTIMIZATIONS_H_
+#define _PRE_NORMALIZE_OPTIMIZATIONS_H_
 
 #include "CallExpr.h"
 #include "symbol.h"
 
 // interface for normalize
-void autoLocalAccess();
+void doPreNormalizeArrayOptimizations();
 
 // interface for resolution
 Expr *preFoldMaybeLocalThis(CallExpr *call);

compiler/optimizations/Makefile.share

@@ -17,7 +17,6 @@
 # limitations under the License.
 
 OPTIMIZATIONS_SRCS = \
-	autoLocalAccess.cpp \
 	bulkCopyRecords.cpp \
 	copyPropagation.cpp \
 	deadCodeElimination.cpp \
@@ -29,6 +28,7 @@ OPTIMIZATIONS_SRCS = \
 	noAliasSets.cpp \
         optimizeForallUnorderedOps.cpp \
 	optimizeOnClauses.cpp \
+	preNormalizeOptimizations.cpp \
 	refPropagation.cpp \
 	remoteValueForwarding.cpp \
 	removeEmptyRecords.cpp \