Add more patterns to scalarize-shapes pass (#3781)

-Adds patterns for propagating shapes through AtenWhereSelf and AtenEqTensor -Adds fold pattern for a rank0 squeezeDim of a full op -Adds support for getting a list from a splat ValueTensorLiteralOp for materializing scalar comparisons in where.self and eq.tensor With a bit of hammering, these changes should unblock several IREE inference failures.
llvm · Oct 11, 2024 · ab62f35 · ab62f35
1 parent 7b11dfc
commit ab62f35
Show file tree

Hide file tree

Showing 2 changed files with 287 additions and 0 deletions.
diff --git a/lib/Dialect/Torch/Transforms/ScalarizeShapes.cpp b/lib/Dialect/Torch/Transforms/ScalarizeShapes.cpp
@@ -63,6 +63,29 @@ LogicalResult getListOperands(Value value, SmallVector<Value> &vals) {
   return success();
 }
 
+LogicalResult constructListFromLiteral(PatternRewriter &rewriter,
+                                       ValueTensorLiteralOp literalOp,
+                                       SmallVector<Value> &vals) {
+  // only supports splat ValueTensorLiterals for now. TODO: add support for
+  // small non-splat valuetensorliterals.
+  auto ty = dyn_cast<ValueTensorType>(literalOp.getType());
+  if (!ty || !ty.hasSizes())
+    return failure();
+  auto attr = dyn_cast_or_null<SplatElementsAttr>(literalOp.getValue());
+  if (!attr)
+    return failure();
+  auto attrInt = dyn_cast<IntegerAttr>(attr.getSplatValue<Attribute>());
+  if (!attrInt)
+    return failure();
+  IntegerType intty = cast<IntegerType>(attrInt.getType());
+  if (!intty.isSignedInteger())
+    return failure();
+  Value materializedVal = rewriter.create<Torch::ConstantIntOp>(
+      literalOp.getLoc(), attrInt.getSInt());
+  vals.resize(vals.size() + ty.getSizes()[0], materializedVal);
+  return success();
+}
+
 LogicalResult getListFromTensor(Value value, SmallVector<Value> &vals) {
   constexpr int64_t kMaxFold = 16;
   if (auto tensor = value.getDefiningOp<Torch::AtenTensorOp>())
@@ -351,6 +374,172 @@ class PropagateAtenSliceTensorPattern
 };
 } // namespace
 
+namespace {
+class PropagateAtenWhereSelfPattern : public OpRewritePattern<AtenWhereSelfOp> {
+public:
+  using OpRewritePattern<AtenWhereSelfOp>::OpRewritePattern;
+  LogicalResult matchAndRewrite(AtenWhereSelfOp op,
+                                PatternRewriter &rewriter) const override {
+    Value condition = op.getCondition();
+    Value self = op.getSelf();
+    Value other = op.getOther();
+    auto conditionTy = dyn_cast<Torch::ValueTensorType>(condition.getType());
+    if (!conditionTy || !conditionTy.hasSizes() ||
+        conditionTy.getSizes().size() != 1)
+      return rewriter.notifyMatchFailure(op, "bad condition type");
+    auto selfTy = dyn_cast<Torch::ValueTensorType>(self.getType());
+    if (!selfTy || !selfTy.hasSizes() || selfTy.getSizes().size() != 1)
+      return rewriter.notifyMatchFailure(op, "bad self type");
+    auto otherTy = dyn_cast<Torch::ValueTensorType>(other.getType());
+    if (!otherTy || !otherTy.hasSizes() || otherTy.getSizes().size() != 1)
+      return rewriter.notifyMatchFailure(op, "bad other type");
+    int64_t conditionSize = selfTy.getSizes()[0];
+    int64_t selfSize = selfTy.getSizes()[0];
+    int64_t otherSize = otherTy.getSizes()[0];
+
+    if (selfSize != otherSize || selfSize != conditionSize)
+      return rewriter.notifyMatchFailure(
+          op,
+          "unimplemented: support for propogating with implicit broadcasting.");
+
+    constexpr int64_t kMaxFold = 16;
+    if (selfSize == Torch::kUnknownSize || selfSize > kMaxFold)
+      return rewriter.notifyMatchFailure(op,
+                                         "arguments are dynamic or too big");
+
+    SmallVector<Value> conditionList, selfList, otherList;
+    if (failed(getListFromTensor(condition, conditionList)) ||
+        (int64_t)conditionList.size() != conditionSize)
+      return failure();
+
+    // If one of these tensors is a value tensor literal op, we will need to
+    // create constant ints in the IR to form a list. Before calling
+    // constructListFromLiteral, we must be certain that the conversion can no
+    // longer fail, otherwise we will cause an infinite loop of creating a
+    // constant and removing it.
+    LogicalResult selfFromList = getListFromTensor(self, selfList);
+    LogicalResult otherFromList = getListFromTensor(other, otherList);
+
+    if (failed(selfFromList) && failed(otherFromList))
+      return rewriter.notifyMatchFailure(
+          op, "At least one operand must succeed at constructing a list");
+
+    auto selfLiteral = self.getDefiningOp<Torch::ValueTensorLiteralOp>();
+    auto otherLiteral = other.getDefiningOp<Torch::ValueTensorLiteralOp>();
+    if (succeeded(selfFromList) && otherLiteral &&
+        failed(constructListFromLiteral(rewriter, otherLiteral, otherList)))
+      return failure();
+    if (succeeded(otherFromList) && selfLiteral &&
+        failed(constructListFromLiteral(rewriter, selfLiteral, selfList)))
+      return failure();
+    if ((int64_t)selfList.size() != selfSize ||
+        (int64_t)otherList.size() != otherSize)
+      // this should only occur if we did not generate IR with
+      // constructListFromLiteral
+      return failure();
+
+    Location loc = op.getLoc();
+    SmallVector<Value> whereVals;
+    auto rank0IntTy = rewriter.getType<Torch::ValueTensorType>(
+        ArrayRef<int64_t>({}), selfTy.getDtype());
+    auto rank0BoolTy = rewriter.getType<Torch::ValueTensorType>(
+        ArrayRef<int64_t>({}), conditionTy.getDtype());
+    for (uint64_t i = 0; i < selfList.size(); i++) {
+      Value rank0Cond = rewriter.create<Torch::PrimNumToTensorScalarOp>(
+          loc, rank0BoolTy, conditionList[i]);
+      Value rank0Self = rewriter.create<Torch::PrimNumToTensorScalarOp>(
+          loc, rank0IntTy, selfList[i]);
+      Value rank0Other = rewriter.create<Torch::PrimNumToTensorScalarOp>(
+          loc, rank0IntTy, otherList[i]);
+      Value rank0Where = rewriter.create<AtenWhereSelfOp>(
+          loc, rank0IntTy, rank0Cond, rank0Self, rank0Other);
+      whereVals.push_back(rewriter.create<AtenItemOp>(
+          loc, rewriter.getType<Torch::IntType>(), rank0Where));
+    }
+    Value list = rewriter.create<Torch::PrimListConstructOp>(
+        op.getLoc(), Torch::ListType::get(whereVals[0].getType()), whereVals);
+    Value cstNone = rewriter.create<Torch::ConstantNoneOp>(op.getLoc());
+    Value cstFalse = rewriter.create<Torch::ConstantBoolOp>(
+        op.getLoc(), rewriter.getBoolAttr(false));
+    rewriter.replaceOpWithNewOp<Torch::AtenTensorOp>(
+        op, op.getType(), list, cstNone, cstNone, cstFalse);
+    return success();
+  }
+};
+} // namespace
+
+namespace {
+class PropagateAtenEqTensorPattern : public OpRewritePattern<AtenEqTensorOp> {
+public:
+  using OpRewritePattern<AtenEqTensorOp>::OpRewritePattern;
+  LogicalResult matchAndRewrite(AtenEqTensorOp op,
+                                PatternRewriter &rewriter) const override {
+    Value self = op.getSelf();
+    Value other = op.getOther();
+    auto selfTy = dyn_cast<Torch::ValueTensorType>(self.getType());
+    if (!selfTy || !selfTy.hasSizes() || selfTy.getSizes().size() != 1)
+      return rewriter.notifyMatchFailure(op, "bad self type");
+    auto otherTy = dyn_cast<Torch::ValueTensorType>(other.getType());
+    if (!otherTy || !otherTy.hasSizes() || otherTy.getSizes().size() != 1)
+      return rewriter.notifyMatchFailure(op, "bad other type");
+    int64_t selfSize = selfTy.getSizes()[0];
+    int64_t otherSize = otherTy.getSizes()[0];
+
+    if (selfSize != otherSize)
+      return rewriter.notifyMatchFailure(
+          op,
+          "unimplemented: support for propogating with implicit broadcasting.");
+
+    constexpr int64_t kMaxFold = 16;
+    if (selfSize == Torch::kUnknownSize || selfSize > kMaxFold ||
+        otherSize == Torch::kUnknownSize || otherSize > kMaxFold)
+      return rewriter.notifyMatchFailure(op,
+                                         "self or other is dynamic or too big");
+
+    SmallVector<Value> selfList, otherList;
+    // If one of these tensors is a value tensor literal op, we will need to
+    // create constant ints in the IR to form a list. Before calling
+    // constructListFromLiteral, we must be certain that the conversion can no
+    // longer fail, otherwise we will cause an infinite loop of creating a
+    // constant and removing it.
+    LogicalResult selfFromList = getListFromTensor(self, selfList);
+    LogicalResult otherFromList = getListFromTensor(other, otherList);
+
+    if (failed(selfFromList) && failed(otherFromList))
+      return rewriter.notifyMatchFailure(
+          op, "At least one operand must succeed at constructing a list");
+
+    auto selfLiteral = self.getDefiningOp<Torch::ValueTensorLiteralOp>();
+    auto otherLiteral = other.getDefiningOp<Torch::ValueTensorLiteralOp>();
+    if (succeeded(selfFromList) && otherLiteral &&
+        failed(constructListFromLiteral(rewriter, otherLiteral, otherList)))
+      return failure();
+    if (succeeded(otherFromList) && selfLiteral &&
+        failed(constructListFromLiteral(rewriter, selfLiteral, selfList)))
+      return failure();
+    if ((int64_t)selfList.size() != selfSize ||
+        (int64_t)otherList.size() != otherSize)
+      // this should only occur if we did not generate IR with
+      // constructListFromLiteral
+      return failure();
+
+    SmallVector<Value> eqVals;
+    for (uint64_t i = 0; i < selfList.size(); i++) {
+      eqVals.push_back(
+          rewriter.create<AtenEqIntOp>(op.getLoc(), selfList[i], otherList[i]));
+    }
+    Value list = rewriter.create<Torch::PrimListConstructOp>(
+        op.getLoc(), Torch::ListType::get(eqVals[0].getType()), eqVals);
+    Value cstNone = rewriter.create<Torch::ConstantNoneOp>(op.getLoc());
+    Value cstFalse = rewriter.create<Torch::ConstantBoolOp>(
+        op.getLoc(), rewriter.getBoolAttr(false));
+    rewriter.replaceOpWithNewOp<Torch::AtenTensorOp>(
+        op, op.getType(), list, cstNone, cstNone, cstFalse);
+    return success();
+  }
+};
+} // namespace
+
 namespace {
 class PropagateAtenItemPattern : public OpRewritePattern<AtenItemOp> {
 public:
@@ -454,6 +643,26 @@ class FoldAtenSqueezePattern : public OpRewritePattern<AtenSqueezeOp> {
 };
 } // namespace
 
+namespace {
+class FoldAtenSqueezeDimPattern : public OpRewritePattern<AtenSqueezeDimOp> {
+public:
+  using OpRewritePattern<AtenSqueezeDimOp>::OpRewritePattern;
+  LogicalResult matchAndRewrite(AtenSqueezeDimOp op,
+                                PatternRewriter &rewriter) const override {
+    auto resultTy = cast<ValueTensorType>(op.getType());
+    if (!resultTy.hasSizes() || resultTy.getSizes().size() != 0)
+      return rewriter.notifyMatchFailure(op, "Unknown result shape");
+
+    if (auto atenFull = op.getSelf().getDefiningOp<AtenFullOp>()) {
+      rewriter.replaceOpWithNewOp<PrimNumToTensorScalarOp>(
+          op, resultTy, atenFull.getFillValue());
+      return success();
+    }
+    return failure();
+  }
+};
+} // namespace
+
 namespace {
 class FoldAtenWhereSelf : public OpRewritePattern<AtenWhereSelfOp> {
 public:
@@ -694,6 +903,8 @@ class ScalarizeShapesPass : public ScalarizeShapesBase<ScalarizeShapesPass> {
                     PropagateAtenSliceTensorPattern, FoldAtenTensorSplatPattern,
                     FoldAtenSqueezePattern, FoldAtenUnsqueezePattern,
                     FoldAtenWhereSelf, CanonicalizeAtenViewPattern,
+                    PropagateAtenEqTensorPattern, PropagateAtenWhereSelfPattern,
+                    FoldAtenSqueezeDimPattern,
                     RemoveUnusedPattern<Torch::AtenIntBoolOp>,
                     RemoveUnusedPattern<Torch::AtenEqIntOp>,
                     RemoveUnusedPattern<Torch::PrimNumToTensorScalarOp>,

diff --git a/test/Dialect/Torch/scalarize-shapes.mlir b/test/Dialect/Torch/scalarize-shapes.mlir
@@ -160,3 +160,79 @@ func.func @unsqueeze_squeeze_combo(%arg0: !torch.vtensor<[?,?,16,64],f32>) -> !t
     %14 = torch.aten.item %13 : !torch.vtensor<[1],si64> -> !torch.int
     return %14 : !torch.int
 }
+
+
+// -----
+
+// CHECK-LABEL: @eq_tensor_and_where_self
+func.func @eq_tensor_and_where_self(%arg0: !torch.vtensor<[?,?],si64>) -> !torch.vtensor<[4],si64> {
+    // CHECK-DAG: %[[false:.*]] = torch.constant.bool false
+    // CHECK-DAG: %[[none:.*]] = torch.constant.none
+    // CHECK-DAG: %[[I1:.*]] = torch.constant.int 1
+    // CHECK-DAG: %[[I0:.*]] = torch.constant.int 0
+    // CHECK-DAG: %[[DIM1:.*]] = torch.aten.size.int %arg0, %[[I1]] : !torch.vtensor<[?,?],si64>, !torch.int -> !torch.int
+    // CHECK-DAG: %[[DIM0:.*]] = torch.aten.size.int %arg0, %[[I0]] : !torch.vtensor<[?,?],si64>, !torch.int -> !torch.int
+    // CHECK: %[[LIST:.*]] = torch.prim.ListConstruct %[[DIM0]], %[[I1]], %[[DIM1]], %[[DIM1]] : (!torch.int, !torch.int, !torch.int, !torch.int) -> !torch.list<int>
+    // CHECK: %[[TENSOR:.*]] = torch.aten.tensor %[[LIST]], %[[none]], %[[none]], %[[false]] : !torch.list<int>, !torch.none, !torch.none, !torch.bool -> !torch.vtensor<[4],si64>
+    // CHECK: return %[[TENSOR]] : !torch.vtensor<[4],si64>
+    %none = torch.constant.none
+    %0 = torch.vtensor.literal(dense<-1> : tensor<4xsi64>) : !torch.vtensor<[4],si64>
+    %1 = torch.vtensor.literal(dense<1> : tensor<4xsi64>) : !torch.vtensor<[4],si64>
+    %false = torch.constant.bool false
+    %int1 = torch.constant.int 1
+    %int0 = torch.constant.int 0
+    %2 = torch.aten.size.int %arg0, %int1 : !torch.vtensor<[?,?],si64>, !torch.int -> !torch.int
+    %3 = torch.aten.size.int %arg0, %int0 : !torch.vtensor<[?,?],si64>, !torch.int -> !torch.int
+    %4 = torch.prim.ListConstruct %3, %int1, %2, %2 : (!torch.int, !torch.int, !torch.int, !torch.int) -> !torch.list<int>
+    %5 = torch.aten.tensor %4, %none, %none, %false : !torch.list<int>, !torch.none, !torch.none, !torch.bool -> !torch.vtensor<[4],si64>
+    %6 = torch.aten.eq.Tensor %5, %0 : !torch.vtensor<[4],si64>, !torch.vtensor<[4],si64> -> !torch.vtensor<[4],i1>
+    %7 = torch.aten.where.self %6, %1, %5 : !torch.vtensor<[4],i1>, !torch.vtensor<[4],si64>, !torch.vtensor<[4],si64> -> !torch.vtensor<[4],si64>
+    return %7 : !torch.vtensor<[4],si64>
+}
+
+
+// -----
+
+// CHECK-LABEL: @eq_tensor_from_tensor_and_literal
+func.func @eq_tensor_from_tensor_and_literal(%arg0: !torch.vtensor<[?,?],si64>) -> !torch.vtensor<[4],i1> {
+    // CHECK-DAG: %[[none:.*]] = torch.constant.none
+    // CHECK-DAG: %[[false:.*]] = torch.constant.bool false
+    // CHECK-DAG: %[[true:.*]] = torch.constant.bool true
+    // CHECK: %[[LIST:.*]] = torch.prim.ListConstruct %[[false]], %[[true]], %[[false]], %[[false]] : (!torch.bool, !torch.bool, !torch.bool, !torch.bool) -> !torch.list<bool>
+    // CHECK: %[[TENSOR:.*]] = torch.aten.tensor %[[LIST]], %[[none]], %[[none]], %[[false]] : !torch.list<bool>, !torch.none, !torch.none, !torch.bool -> !torch.vtensor<[4],i1>
+    // CHECK: return %[[TENSOR]] : !torch.vtensor<[4],i1>
+    %none = torch.constant.none
+    %0 = torch.vtensor.literal(dense<-1> : tensor<4xsi64>) : !torch.vtensor<[4],si64>
+    %1 = torch.vtensor.literal(dense<1> : tensor<4xsi64>) : !torch.vtensor<[4],si64>
+    %false = torch.constant.bool false
+    %int1 = torch.constant.int 1
+    %int-1 = torch.constant.int -1
+    %int0 = torch.constant.int 0
+    %2 = torch.aten.size.int %arg0, %int1 : !torch.vtensor<[?,?],si64>, !torch.int -> !torch.int
+    %3 = torch.aten.size.int %arg0, %int0 : !torch.vtensor<[?,?],si64>, !torch.int -> !torch.int
+    %4 = torch.prim.ListConstruct %3, %int-1, %2, %2 : (!torch.int, !torch.int, !torch.int, !torch.int) -> !torch.list<int>
+    %5 = torch.aten.tensor %4, %none, %none, %false : !torch.list<int>, !torch.none, !torch.none, !torch.bool -> !torch.vtensor<[4],si64>
+    %6 = torch.aten.eq.Tensor %5, %0 : !torch.vtensor<[4],si64>, !torch.vtensor<[4],si64> -> !torch.vtensor<[4],i1>
+    return %6 : !torch.vtensor<[4],i1>
+}
+
+
+
+// -----
+
+// CHECK-LABEL: @squeeze_dim_full_fold
+func.func @squeeze_dim_full_fold(%arg0: !torch.vtensor<[?,?],si64>) -> !torch.int {
+    // CHECK: %[[I0:.*]] = torch.constant.int 0
+    // CHECK: %[[SZE:.*]] = torch.aten.size.int %arg0, %[[I0]] : !torch.vtensor<[?,?],si64>, !torch.int -> !torch.int
+    // CHECK: return %[[SZE]] : !torch.int
+    %int0 = torch.constant.int 0
+    %int1 = torch.constant.int 1
+    %none = torch.constant.none
+    %false = torch.constant.bool false
+    %51 = torch.aten.size.int %arg0, %int0 : !torch.vtensor<[?,?],si64>, !torch.int -> !torch.int
+    %55 = torch.prim.ListConstruct %int1 : (!torch.int) -> !torch.list<int>
+    %56 = torch.aten.full %55, %51, %none, %none, %none, %false : !torch.list<int>, !torch.int, !torch.none, !torch.none, !torch.none, !torch.bool -> !torch.vtensor<[1],si64>
+    %57 = torch.aten.squeeze.dim %56, %int0 : !torch.vtensor<[1],si64>, !torch.int -> !torch.vtensor<[],si64>
+    %58 = torch.aten.item %57 : !torch.vtensor<[],si64> -> !torch.int
+    return %58 : !torch.int
+}