[Unroll/UnrollAndJam/Vectorizer/Distribute] Add followup loop attributes.

When multiple loop transformation are defined in a loop's metadata, their order of execution is defined by the order of their respective passes in the pass pipeline. For instance, e.g.

    #pragma clang loop unroll_and_jam(enable)
    #pragma clang loop distribute(enable)

is the same as

    #pragma clang loop distribute(enable)
    #pragma clang loop unroll_and_jam(enable)

and will try to loop-distribute before Unroll-And-Jam because the LoopDistribute pass is scheduled after UnrollAndJam pass. UnrollAndJamPass only supports one inner loop, i.e. it will necessarily fail after loop distribution. It is not possible to specify another execution order. Also,t the order of passes in the pipeline is subject to change between versions of LLVM, optimization options and which pass manager is used.

This patch adds 'followup' attributes to various loop transformation passes. These attributes define which attributes the resulting loop of a transformation should have. For instance,

    !0 = !{!0, !1, !2}
    !1 = !{!"llvm.loop.unroll_and_jam.enable"}
    !2 = !{!"llvm.loop.unroll_and_jam.followup_inner", !3}
    !3 = !{!"llvm.loop.distribute.enable"}

defines a loop ID (!0) to be unrolled-and-jammed (!1) and then the attribute !3 to be added to the jammed inner loop, which contains the instruction to distribute the inner loop.

Currently, in both pass managers, pass execution is in a fixed order and UnrollAndJamPass will not execute again after LoopDistribute. We hope to fix this in the future by allowing pass managers to run passes until a fixpoint is reached, use Polly to perform these transformations, or add a loop transformation pass which takes the order issue into account.

For mandatory/forced transformations (e.g. by having been declared by #pragma omp simd), the user must be notified when a transformation could not be performed. It is not possible that the responsible pass emits such a warning because the transformation might be 'hidden' in a followup attribute when it is executed, or it is not present in the pipeline at all. For this reason, this patche introduces a WarnMissedTransformations pass, to warn about orphaned transformations.

Since this changes the user-visible diagnostic message when a transformation is applied, two test cases in the clang repository need to be updated.

To ensure that no other transformation is executed before the intended one, the attribute `llvm.loop.disable_nonforced` can be added which should disable transformation heuristics before the intended transformation is applied. E.g. it would be surprising if a loop is distributed before a #pragma unroll_and_jam is applied.

With more supported code transformations (loop fusion, interchange, stripmining, offloading, etc.), transformations can be used as building blocks for more complex transformations (e.g. stripmining+stripmining+interchange -> tiling).

Reviewed By: hfinkel, dmgreen

Differential Revision: https://reviews.llvm.org/D49281
Differential Revision: https://reviews.llvm.org/D55288

llvm-svn: 348944

1 files changed, 66 insertions, 11 deletions

diff --git a/llvm/lib/Transforms/Scalar/LoopUnrollAndJamPass.cpp b/llvm/lib/Transforms/Scalar/LoopUnrollAndJamPass.cpp
index 30dfb9b5dd2..da46210b6fd 100644
--- a/llvm/lib/Transforms/Scalar/LoopUnrollAndJamPass.cpp
+++ b/llvm/lib/Transforms/Scalar/LoopUnrollAndJamPass.cpp
@@ -56,6 +56,20 @@ using namespace llvm;
 
 #define DEBUG_TYPE "loop-unroll-and-jam"
 
+/// @{
+/// Metadata attribute names
+static const char *const LLVMLoopUnrollAndJamFollowupAll =
+    "llvm.loop.unroll_and_jam.followup_all";
+static const char *const LLVMLoopUnrollAndJamFollowupInner =
+    "llvm.loop.unroll_and_jam.followup_inner";
+static const char *const LLVMLoopUnrollAndJamFollowupOuter =
+    "llvm.loop.unroll_and_jam.followup_outer";
+static const char *const LLVMLoopUnrollAndJamFollowupRemainderInner =
+    "llvm.loop.unroll_and_jam.followup_remainder_inner";
+static const char *const LLVMLoopUnrollAndJamFollowupRemainderOuter =
+    "llvm.loop.unroll_and_jam.followup_remainder_outer";
+/// @}
+
 static cl::opt<bool>
     AllowUnrollAndJam("allow-unroll-and-jam", cl::Hidden,
                       cl::desc("Allows loops to be unroll-and-jammed."));
@@ -112,11 +126,6 @@ static bool HasUnrollAndJamEnablePragma(const Loop *L) {
   return GetUnrollMetadataForLoop(L, "llvm.loop.unroll_and_jam.enable");
 }
 
-// Returns true if the loop has an unroll_and_jam(disable) pragma.
-static bool HasUnrollAndJamDisablePragma(const Loop *L) {
-  return GetUnrollMetadataForLoop(L, "llvm.loop.unroll_and_jam.disable");
-}
-
 // If loop has an unroll_and_jam_count pragma return the (necessarily
 // positive) value from the pragma.  Otherwise return 0.
 static unsigned UnrollAndJamCountPragmaValue(const Loop *L) {
@@ -299,13 +308,16 @@ tryToUnrollAndJamLoop(Loop *L, DominatorTree &DT, LoopInfo *LI,
                     << L->getHeader()->getParent()->getName() << "] Loop %"
                     << L->getHeader()->getName() << "\n");
 
+  TransformationMode EnableMode = hasUnrollAndJamTransformation(L);
+  if (EnableMode & TM_Disable)
+    return LoopUnrollResult::Unmodified;
+
   // A loop with any unroll pragma (enabling/disabling/count/etc) is left for
   // the unroller, so long as it does not explicitly have unroll_and_jam
   // metadata. This means #pragma nounroll will disable unroll and jam as well
   // as unrolling
-  if (HasUnrollAndJamDisablePragma(L) ||
-      (HasAnyUnrollPragma(L, "llvm.loop.unroll.") &&
-       !HasAnyUnrollPragma(L, "llvm.loop.unroll_and_jam."))) {
+  if (HasAnyUnrollPragma(L, "llvm.loop.unroll.") &&
+      !HasAnyUnrollPragma(L, "llvm.loop.unroll_and_jam.")) {
     LLVM_DEBUG(dbgs() << "  Disabled due to pragma.\n");
     return LoopUnrollResult::Unmodified;
   }
@@ -344,6 +356,19 @@ tryToUnrollAndJamLoop(Loop *L, DominatorTree &DT, LoopInfo *LI,
     return LoopUnrollResult::Unmodified;
   }
 
+  // Save original loop IDs for after the transformation.
+  MDNode *OrigOuterLoopID = L->getLoopID();
+  MDNode *OrigSubLoopID = SubLoop->getLoopID();
+
+  // To assign the loop id of the epilogue, assign it before unrolling it so it
+  // is applied to every inner loop of the epilogue. We later apply the loop ID
+  // for the jammed inner loop.
+  Optional<MDNode *> NewInnerEpilogueLoopID = makeFollowupLoopID(
+      OrigOuterLoopID, {LLVMLoopUnrollAndJamFollowupAll,
+                        LLVMLoopUnrollAndJamFollowupRemainderInner});
+  if (NewInnerEpilogueLoopID.hasValue())
+    SubLoop->setLoopID(NewInnerEpilogueLoopID.getValue());
+
   // Find trip count and trip multiple
   unsigned OuterTripCount = SE.getSmallConstantTripCount(L, Latch);
   unsigned OuterTripMultiple = SE.getSmallConstantTripMultiple(L, Latch);
@@ -359,9 +384,39 @@ tryToUnrollAndJamLoop(Loop *L, DominatorTree &DT, LoopInfo *LI,
   if (OuterTripCount && UP.Count > OuterTripCount)
     UP.Count = OuterTripCount;
 
-  LoopUnrollResult UnrollResult =
-      UnrollAndJamLoop(L, UP.Count, OuterTripCount, OuterTripMultiple,
-                       UP.UnrollRemainder, LI, &SE, &DT, &AC, &ORE);
+  Loop *EpilogueOuterLoop = nullptr;
+  LoopUnrollResult UnrollResult = UnrollAndJamLoop(
+      L, UP.Count, OuterTripCount, OuterTripMultiple, UP.UnrollRemainder, LI,
+      &SE, &DT, &AC, &ORE, &EpilogueOuterLoop);
+
+  // Assign new loop attributes.
+  if (EpilogueOuterLoop) {
+    Optional<MDNode *> NewOuterEpilogueLoopID = makeFollowupLoopID(
+        OrigOuterLoopID, {LLVMLoopUnrollAndJamFollowupAll,
+                          LLVMLoopUnrollAndJamFollowupRemainderOuter});
+    if (NewOuterEpilogueLoopID.hasValue())
+      EpilogueOuterLoop->setLoopID(NewOuterEpilogueLoopID.getValue());
+  }
+
+  Optional<MDNode *> NewInnerLoopID =
+      makeFollowupLoopID(OrigOuterLoopID, {LLVMLoopUnrollAndJamFollowupAll,
+                                           LLVMLoopUnrollAndJamFollowupInner});
+  if (NewInnerLoopID.hasValue())
+    SubLoop->setLoopID(NewInnerLoopID.getValue());
+  else
+    SubLoop->setLoopID(OrigSubLoopID);
+
+  if (UnrollResult == LoopUnrollResult::PartiallyUnrolled) {
+    Optional<MDNode *> NewOuterLoopID = makeFollowupLoopID(
+        OrigOuterLoopID,
+        {LLVMLoopUnrollAndJamFollowupAll, LLVMLoopUnrollAndJamFollowupOuter});
+    if (NewOuterLoopID.hasValue()) {
+      L->setLoopID(NewOuterLoopID.getValue());
+
+      // Do not setLoopAlreadyUnrolled if a followup was given.
+      return UnrollResult;
+    }
+  }
 
   // If loop has an unroll count pragma or unrolled by explicitly set count
   // mark loop as unrolled to prevent unrolling beyond that requested.