Fix merge conflicts.

NVIDIA · Jan 20, 2025 · 9163101 · 9163101
1 parent c5939f0
commit 9163101
Show file tree

Hide file tree

Showing 3 changed files with 28 additions and 15 deletions.
diff --git a/csrc/host_ir/container.cpp b/csrc/host_ir/container.cpp
@@ -14,6 +14,7 @@
 #include <ir/utils.h>
 #include <kernel_ir.h>
 #include <ops/all_ops.h>
+#include <runtime/executor.h>
 
 namespace nvfuser {
 
@@ -32,6 +33,24 @@ std::ostream& HostIrContainer::print(std::ostream& os) const {
   return os;
 }
 
+const std::vector<Expr*>& HostIrContainer::topLevelExprs() const {
+  return top_level_exprs_;
+}
+
+void HostIrContainer::pushBackTopLevelExprs(Expr* expr) {
+  assertInContainer(expr, "Cannot add expr, ");
+  return top_level_exprs_.push_back(expr);
+}
+
+void HostIrContainer::pushBackKernelExecutor(
+    std::unique_ptr<KernelExecutor> ke) {
+  return kernel_executors_.push_back(std::move(ke));
+}
+
+KernelExecutor* HostIrContainer::getKernelExecutor(int64_t index) const {
+  return kernel_executors_.at(index).get();
+}
+
 } // namespace hir
 
 } // namespace nvfuser
diff --git a/csrc/host_ir/container.h b/csrc/host_ir/container.h
@@ -9,10 +9,11 @@
 
 #include <fusion.h>
 #include <host_ir/host_ir.h>
-#include <runtime/executor.h>
 
 namespace nvfuser {
 
+class KernelExecutor;
+
 namespace hir {
 
 /*
@@ -33,22 +34,13 @@ class HostIrContainer final : public Fusion {
   //! Print to an output stream
   std::ostream& print(std::ostream& os) const;
 
-  const auto& topLevelExprs() const {
-    return top_level_exprs_;
-  }
+  const std::vector<Expr*>& topLevelExprs() const;
 
-  void pushBackTopLevelExprs(Expr* expr) {
-    assertInContainer(expr, "Cannot add expr, ");
-    return top_level_exprs_.push_back(expr);
-  }
+  void pushBackTopLevelExprs(Expr* expr);
 
-  void pushBackKernelExecutor(std::unique_ptr<KernelExecutor> ke) {
-    return kernel_executors_.push_back(std::move(ke));
-  }
+  void pushBackKernelExecutor(std::unique_ptr<KernelExecutor> ke);
 
-  KernelExecutor* getKernelExecutor(int64_t index) const {
-    return kernel_executors_.at(index).get();
-  }
+  KernelExecutor* getKernelExecutor(int64_t index) const;
 
   Stream* getDefaultStream();
 

diff --git a/tests/cpp/test_rope.cpp b/tests/cpp/test_rope.cpp
@@ -924,7 +924,9 @@ TEST_F(RopeTest, EndingRepeat) {
       runtime->schedulerHeuristics()->heuristicsList().front();
   EXPECT_EQ(heuristic_param->scheduler_type, SchedulerType::Resize);
   Fusion* scheduled_fusion =
-      runtime->executors().at(0)->as<KernelExecutor>()->kernel();
+      dynamic_cast<KernelExecutor*>(runtime->executors().at(0).get())
+          ->compiledKernel()
+          ->fusion();
 
   // Check the loop domain of the reference. It should look like:
   //