diff --git a/third_party/nvfuser/CMakeLists.txt b/third_party/nvfuser/CMakeLists.txt
index 1cca6680aebd0..68133e7e2c8fa 100644
--- a/third_party/nvfuser/CMakeLists.txt
+++ b/third_party/nvfuser/CMakeLists.txt
@@ -315,6 +315,7 @@ if(USE_CUDA)
   list(APPEND JIT_TEST_SRCS ${NVFUSER_ROOT}/test/test_gpu_indexing_ops.cpp)
   list(APPEND JIT_TEST_SRCS ${NVFUSER_ROOT}/test/test_gpu_indexing.cpp)
   list(APPEND JIT_TEST_SRCS ${NVFUSER_ROOT}/test/test_gpu_gather_ops.cpp)
+  list(APPEND JIT_TEST_SRCS ${NVFUSER_ROOT}/test/test_gpu_match_frontend.cpp)
 
   add_executable(${NVFUSER_TESTS}
              ${TORCH_ROOT}/test/cpp/common/main.cpp
diff --git a/third_party/nvfuser/test/test_gpu_match_frontend.cpp b/third_party/nvfuser/test/test_gpu_match_frontend.cpp
index 2022d91e4ffa9..e82216a973342 100644
--- a/third_party/nvfuser/test/test_gpu_match_frontend.cpp
+++ b/third_party/nvfuser/test/test_gpu_match_frontend.cpp
@@ -6,18 +6,18 @@
 #include <gmock/gmock-matchers.h>
 #include <gtest/gtest.h>
 
-#include <torch/csrc/jit/codegen/cuda/arith.h>
-#include <torch/csrc/jit/codegen/cuda/executor.h>
-#include <torch/csrc/jit/codegen/cuda/inlining.h>
-#include <torch/csrc/jit/codegen/cuda/ir_all_nodes.h>
-#include <torch/csrc/jit/codegen/cuda/ir_builder.h>
-#include <torch/csrc/jit/codegen/cuda/kernel_cache.h>
-#include <torch/csrc/jit/codegen/cuda/scheduler/all_schedulers.h>
-#include <torch/csrc/jit/codegen/cuda/scheduler/utils.h>
+#include <arith.h>
+#include <executor.h>
+#include <inlining.h>
+#include <ir_all_nodes.h>
+#include <ir_builder.h>
+#include <kernel_cache.h>
+#include <scheduler/all_schedulers.h>
+#include <scheduler/utils.h>
 
 #include <test/cpp/jit/test_utils.h>
-#include <torch/csrc/jit/codegen/cuda/test/test_gpu_validator.h>
-#include <torch/csrc/jit/codegen/cuda/test/test_utils.h>
+#include <test/test_gpu_validator.h>
+#include <test/test_utils.h>
 
 #include <torch/torch.h>