Resolve conflicts by recomputation

csrc/scheduler/resize.cpp

@@ -73,19 +73,6 @@ bool ResizeScheduler::canScheduleCompileTime(Fusion* fusion) {
 
   auto resize_based_tensor_ops = ir_utils::getOpsOfType<SliceOp, PadOp>(fusion);
 
-  if (auto non_exclusive_resizes = scheduler_tools::getNonExclusiveResizeInfo(
-          resize_based_tensor_ops, id_model.idGraph(IdMappingMode::EXACT));
-      !non_exclusive_resizes.empty()) {
-    std::stringstream msg;
-    msg << "Propagation of resizes would affect fusion outputs.";
-    for (const auto& [tv, resize_ids] : non_exclusive_resizes) {
-      msg << " Resize input tv: " << tv->toString()
-          << ", resize input ID groups: " << nvfuser::toString(resize_ids);
-    }
-    scheduler_debug_utils::canScheduleRejectReason(schedulerType(), msg.str());
-    return false;
-  }
-
   // Slicing of or to a broadcast ID is not allowed yet.
   for (auto tensor_op : resize_based_tensor_ops) {
     TensorView* out_tv = tensor_op->output(0)->as<TensorView>();
@@ -133,6 +120,30 @@ bool ResizeScheduler::canScheduleCompileTime(Fusion* fusion) {
     return false;
   }
 
+  for (auto out_tv : ir_utils::filterByType<TensorView>(fusion->outputs())) {
+    if (out_tv == ref_tv) {
+      continue;
+    }
+    auto exprs = ValGraphBFS::getExprGroupsBetween(
+                     broadcast_graph,
+                     broadcast_graph.toGroups(ref_tv->getLogicalDomain()),
+                     broadcast_graph.toGroups(out_tv->getLogicalDomain()),
+                     /*require_all_to_visited=*/false)
+                     .first;
+    for (const auto& [expr_g, dir] : exprs) {
+      if (expr_g->front()->isA<Resize>()) {
+        std::stringstream msg;
+        msg << "Resize between reference and output not allowed.";
+        msg << " Reference: " << ref_tv->toString()
+            << ". Output: " << out_tv->toString()
+            << ". Resize: " << expr_g->front()->toString();
+        scheduler_debug_utils::canScheduleRejectReason(
+            schedulerType(), msg.str());
+        return false;
+      }
+    }
+  }
+
   // Disable the scheduler if there's a squeeze op. The loop option
   // may also need to be enabled in that case, but that option is not
   // turned on automatically yet.
@@ -163,6 +174,21 @@ void ResizeScheduler::schedule(Fusion* fusion, const HeuristicParams* params) {
   scheduler_utils::cacheInputs(fusion, true);
   scheduler_utils::cacheAndForkOutputs(fusion, true);
 
+  auto resize_based_tensor_ops = ir_utils::getOpsOfType<SliceOp, PadOp>(fusion);
+
+  IdModel id_model(fusion, /*build_graphs=*/false);
+  const auto& exact_graph = id_model.buildExactGraph();
+
+  // Replicate resize inputs if necessary to avoid conflicting propagations
+  for (const auto& [out_tv, exlusivity_info] :
+       scheduler_tools::getNonExclusiveResizeInfo(
+           resize_based_tensor_ops, exact_graph)) {
+    auto resize_based_op = out_tv->definition();
+    auto inp_tv = resize_based_op->input(0)->as<TensorView>();
+    auto inp_tv_copy = RecomputeTv::recompute(inp_tv);
+    ir_utils::replaceValInExprInputs(resize_based_op, inp_tv, inp_tv_copy);
+  }
+
   for (auto expr : fusion->exprs()) {
     if (!expr->isOneOf<SliceOp, PadOp>()) {
       continue;
@@ -186,9 +212,14 @@ void ResizeScheduler::schedule(Fusion* fusion, const HeuristicParams* params) {
   ref_tv->axis(-1)->parallelize(ParallelType::TIDx);
   ref_tv->axis(-2)->parallelize(ParallelType::BIDx);
 
-  // Propagate the reference to the other tensors
+  // Propagate the reference to the other tensors. Note that the
+  // update flag is enabled so to workaround the resize propagation
+  // issue. This may not work if there's a tensor that is reshaped
+  // from the reference tensor, but that should not be the case as the
+  // reference is picked by the same routine used for the pointwise
+  // scheduler.
   scheduler_tools::scheduleLoopDomainsLike(
-      fusion->allTvs(), ref_tv->getLoopDomain());
+      fusion->allTvs(), ref_tv->getLoopDomain(), true);
 
   inlineMost();
 

csrc/scheduler/tools/resize_utils.cpp

@@ -66,13 +66,14 @@ void propagateResizeToInputs(Expr* resize_tensor_op) {
   }
 }
 
-std::unordered_map<TensorView*, ValGroups> getNonExclusiveResizeInfo(
+std::unordered_map<TensorView*, ResizeExclusivityInfo> getNonExclusiveResizeInfo(
     const std::vector<Expr*>& ordered_resize_tensor_ops,
-    const ValGraph& exact_graph) {
+    const ValGraph& exact_graph,
+    bool ignore_fusion_inputs) {
   NVF_ERROR(!ordered_resize_tensor_ops.empty());
   Fusion* fusion = ordered_resize_tensor_ops[0]->fusion();
 
-  std::unordered_map<TensorView*, ValGroups> non_exclusive_resizes;
+  std::unordered_map<TensorView*, ResizeExclusivityInfo> non_exclusive_resizes;
 
   std::unordered_set<Val*> inputs{
       fusion->inputs().begin(), fusion->inputs().end()};
@@ -95,6 +96,8 @@ std::unordered_map<TensorView*, ValGroups> getNonExclusiveResizeInfo(
     auto inp_tv = dynamic_cast<TensorView*>(resize_tensor_op->inputs().at(0));
     auto out_tv = dynamic_cast<TensorView*>(resize_tensor_op->outputs().at(0));
 
+    ResizeExclusivityInfo info;
+
     ValGroups resize_inp_ids = get_root_to_logical_resizes(out_tv);
     NVF_ERROR(!resize_inp_ids.empty());
 
@@ -107,6 +110,10 @@ std::unordered_map<TensorView*, ValGroups> getNonExclusiveResizeInfo(
     // visible changes through the tensor, the resize is considered
     // non-exclusive.
     for (auto dep_tv : ir_utils::filterByType<TensorView>(dep_vals)) {
+      if (ignore_fusion_inputs && dep_tv->isFusionInput()) {
+        continue;
+      }
+
       bool maybe_non_exclusive = false;
 
       if (dep_tv->isFusionOutput()) {
@@ -159,10 +166,16 @@ std::unordered_map<TensorView*, ValGroups> getNonExclusiveResizeInfo(
         }
 
         // This resize input ID is not exclusively used
-        non_exclusive_resizes[inp_tv].pushBack(resize_inp_id);
+        // non_exclusive_resizes[inp_tv].first.pushBack(resize_inp_id);
+        info.shared_tvs.push_back(dep_tv);
+        info.resized_ids.pushBack(resize_inp_id);
       }
     }
 
+    if (!info.shared_tvs.empty()) {
+      NVF_ERROR(non_exclusive_resizes.emplace(out_tv, info).second);
+    }
+
     // Analysis of exclusiveness until in_tv is done. Following
     // resize-based tensor ops do not need to check the same section
     // of the fusion and can start from out_tv.

csrc/scheduler/tools/resize_utils.h

@@ -94,9 +94,24 @@ void propagateResizeToInputs(Expr* resize_op);
 // The function returns a map from tensors that are input to
 // non-exclusive ops to their resize input ID groups. This map will be
 // used to resolve the non-exclusiveness by replication.
-std::unordered_map<TensorView*, ValGroups> getNonExclusiveResizeInfo(
+struct ResizeExclusivityInfo {
+  std::vector<TensorView*> shared_tvs;
+  // std::unordered_map<TensorView*, ValGroups> resized_ids;
+  ValGroups resized_ids;
+
+  bool operator==(const ResizeExclusivityInfo& other) const {
+    return shared_tvs == other.shared_tvs && resized_ids == other.resized_ids;
+  }
+
+  bool operator!=(const ResizeExclusivityInfo& other) const {
+    return !(*this == other);
+  }
+};
+
+std::unordered_map<TensorView*, ResizeExclusivityInfo> getNonExclusiveResizeInfo(
     const std::vector<Expr*>& ordered_resize_tensor_ops,
-    const ValGraph& exact_graph);
+    const ValGraph& exact_graph,
+    bool ignore_fusion_inputs = false);
 
 } // namespace scheduler_tools
 } // namespace nvfuser