diff --git a/benchmarks/DeepLearning/Ops/CMakeLists.txt b/benchmarks/DeepLearning/Ops/CMakeLists.txt
index 2978cc80..d989d8d4 100644
--- a/benchmarks/DeepLearning/Ops/CMakeLists.txt
+++ b/benchmarks/DeepLearning/Ops/CMakeLists.txt
@@ -14,4 +14,5 @@ add_subdirectory(BatchMatMulOp)
 add_subdirectory(MathExpOp)
 add_subdirectory(ReduceAddfOp)
 add_subdirectory(ReduceMaxfOp)
-add_subdirectory(SoftmaxExpSumDivOp)
\ No newline at end of file
+add_subdirectory(SoftmaxExpSumDivOp)
+add_subdirectory(RFFTOp)
diff --git a/benchmarks/DeepLearning/Ops/RFFTOp/CMakeLists.txt b/benchmarks/DeepLearning/Ops/RFFTOp/CMakeLists.txt
new file mode 100644
index 00000000..35a81578
--- /dev/null
+++ b/benchmarks/DeepLearning/Ops/RFFTOp/CMakeLists.txt
@@ -0,0 +1,46 @@
+cmake_minimum_required(VERSION 3.10)
+project(RFFTBenchmark)
+
+add_custom_command(
+  OUTPUT
+    ${CMAKE_CURRENT_SOURCE_DIR}/result.txt
+  COMMAND
+    python3 ${CMAKE_CURRENT_SOURCE_DIR}/RFFT.py ${CMAKE_CURRENT_SOURCE_DIR}
+  COMMENT "Generating result.txt"
+)
+
+add_custom_target(generate_result_txt ALL
+  DEPENDS ${CMAKE_CURRENT_SOURCE_DIR}/result.txt
+  COMMENT "Ensure result.txt is generated"
+)
+
+# Define variables for the cross-compilation toolchain and options.
+if (CROSS_COMPILE_RVV)
+  set(RISCV_GNU_TOOLCHAIN ${BUDDY_MLIR_BUILD_DIR}/thirdparty/riscv-gnu-toolchain)
+  set(RISCV_GNU_TOOLCHAIN_SYSROOT ${RISCV_GNU_TOOLCHAIN}/sysroot)
+  set(BUDDY_OPT_TRIPLE riscv64)
+  set(BUDDY_OPT_ATTR +v,+m)
+endif()
+
+# Add the executable for the benchmark
+add_executable(dl-op-linalg-RFFT-benchmark
+  GoogleBenchmarkMain.cpp
+)
+
+# Define the benchmark tool
+set(BenchmarkTool GoogleBenchmark)
+
+target_link_directories(dl-op-linalg-RFFT-benchmark PRIVATE 
+  ${BUDDY_MLIR_LIB_DIR}
+  ${LLVM_MLIR_LIBRARY_DIR}
+)
+
+# Link the benchmark executable with the necessary libraries
+target_link_libraries(dl-op-linalg-RFFT-benchmark PRIVATE
+  ${BenchmarkTool}
+  BuddyLibDAP
+  mlir_c_runner_utils
+  GoogleBenchmark
+)
+
+add_dependencies(dl-op-linalg-RFFT-benchmark generate_result_txt)
diff --git a/benchmarks/DeepLearning/Ops/RFFTOp/GoogleBenchmarkMain.cpp b/benchmarks/DeepLearning/Ops/RFFTOp/GoogleBenchmarkMain.cpp
new file mode 100644
index 00000000..0811fc83
--- /dev/null
+++ b/benchmarks/DeepLearning/Ops/RFFTOp/GoogleBenchmarkMain.cpp
@@ -0,0 +1,120 @@
+//===- GoogleBenchmarkMain.cpp---------------------------------------------===//
+//
+// Licensed under the Apache License, Version 2.0 (the "License");
+// you may not use this file except in compliance with the License.
+// You may obtain a copy of the License at
+//
+//     http://www.apache.org/licenses/LICENSE-2.0
+//
+// Unless required by applicable law or agreed to in writing, software
+// distributed under the License is distributed on an "AS IS" BASIS,
+// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+// See the License for the specific language governing permissions and
+// limitations under the License.
+//
+//===----------------------------------------------------------------------===//
+//
+// This file implements the benchmark for RFFT operation.
+//
+//===----------------------------------------------------------------------===//
+
+#include <benchmark/benchmark.h>
+#include <buddy/Core/Container.h>
+#include <buddy/DAP/DAP.h>
+#include <iostream>
+#include <random>
+
+#define testLength 20
+
+// Helper functions and variables.
+namespace {
+const std::string PASS = "\033[32mPASS\033[0m";
+const std::string FAIL = "\033[31mFAIL\033[0m";
+
+bool areArraysEqual(float array1[], float array2[], int size) {
+  for (int i = 0; i < size; ++i) {
+    if (array1[i] != array2[i]) {
+      return false;
+    }
+  }
+  return true;
+}
+double *inputAlign0 = new double[testLength];
+intptr_t inputSizes0[1] = {testLength};
+
+void initializeInputAlign0() {
+
+  for (int i = 0; i < testLength; ++i) {
+    inputAlign0[i] = static_cast<double>(i);
+  }
+}
+
+} // namespace
+
+static void BUDDY_RFFT(benchmark::State &state) {
+  MemRef<double, 1> inputMemRef0(inputAlign0, inputSizes0);
+  for (auto _ : state) {
+    for (int i = 0; i < state.range(0); ++i) {
+      dap::RFFT(&inputMemRef0);
+    }
+  }
+}
+
+BENCHMARK(BUDDY_RFFT)->Arg(1)->Unit(benchmark::kMillisecond);
+
+void verification() {
+
+  std::vector<double> fileData;
+  std::ifstream inputFile(
+      "../../benchmarks/DeepLearning/Ops/RFFTOp/result.txt");
+
+  double value;
+  while (inputFile >> value) {
+    fileData.push_back(value);
+  }
+  inputFile.close();
+
+  double *inputAlign = new double[testLength];
+  for (int i = 0; i < testLength; ++i) {
+    inputAlign[i] = static_cast<double>(i);
+  }
+  intptr_t inputSizes[1] = {testLength};
+  MemRef<double, 1> inputMemRef(inputAlign, inputSizes);
+
+  dap::RFFT(&inputMemRef);
+
+  // Get the result array.
+  auto resultRFFT = inputMemRef.getData();
+
+  std::cout << "Length : " << fileData.size() << std::endl;
+
+  bool isEqual = true;
+  double tolerance = 1e-2;
+  size_t minSize = fileData.size();
+  for (size_t i = 0; i < minSize; ++i) {
+    if (std::abs(resultRFFT[i] - fileData[i]) > tolerance) {
+      isEqual = false;
+    }
+  }
+
+  std::cout << "-----------------------------------------------------------"
+            << std::endl;
+  std::cout << "Correctness Verification: " << (isEqual ? PASS : FAIL)
+            << std::endl;
+  std::cout << "-----------------------------------------------------------"
+            << std::endl;
+
+}
+
+int main(int argc, char **argv) {
+  // Run benchmark.
+  initializeInputAlign0();
+
+  ::benchmark::Initialize(&argc, argv);
+
+  ::benchmark::RunSpecifiedBenchmarks();
+
+  // Run correctness verification.
+  verification();
+  return 0;
+}
diff --git a/benchmarks/DeepLearning/Ops/RFFTOp/RFFT.py b/benchmarks/DeepLearning/Ops/RFFTOp/RFFT.py
new file mode 100644
index 00000000..ee7b1454
--- /dev/null
+++ b/benchmarks/DeepLearning/Ops/RFFTOp/RFFT.py
@@ -0,0 +1,48 @@
+import numpy as np
+import sys
+import os
+import timeit
+
+if len(sys.argv) != 2:
+    print("need file path")
+    sys.exit(1)
+
+output_path = sys.argv[1]
+output_file = os.path.join(output_path, 'result.txt')
+time_output_file = os.path.join(output_path, 'time_result.txt')
+
+if os.path.exists(output_file):
+    os.remove(output_file)
+if os.path.exists(time_output_file):
+    os.remove(time_output_file)
+
+def compute_rfft():
+    arr = np.arange(0, 20)
+    rfft_result = np.fft.rfft(arr)
+    output_lines = []
+    first_value = rfft_result[0].real
+    output_lines.append(f"{first_value:.0f}")
+    for c in rfft_result[1:]:
+        real_part = c.real
+        imag_part = c.imag
+        output_lines.append(f"{real_part:.0f}")
+        output_lines.append(f"{imag_part:.2f}")
+    return output_lines
+
+# set a timer with timeit
+execution_time = timeit.timeit(compute_rfft, number=1) * 1000
+
+# get RFFT result
+rfft_result_lines = compute_rfft()
+
+# record the result
+with open(output_file, 'w') as f:
+    for line in rfft_result_lines:
+        f.write(line + '\n')
+
+with open(time_output_file, 'w') as f_time:
+    f_time.write(f"Execution time for RFFT: {execution_time:.6f} milliseconds\n")
+
+
+print(f"RFFT result saved to '{output_file}'")
+print(f"Execution time saved to '{time_output_file}'")
diff --git a/benchmarks/DeepLearning/README.md b/benchmarks/DeepLearning/README.md
index 8c643e71..0e5a25d1 100644
--- a/benchmarks/DeepLearning/README.md
+++ b/benchmarks/DeepLearning/README.md
@@ -39,6 +39,7 @@ The table below lists the benchmark cases at the operation level.
 | Reduce Addf | `ninja dl-op-reduce-addf-benchmark` | This benchmark evaluates optimization strategies for the `reduce.addf` operation. The benchmark size can be adjusted in [this file](./Ops/ReduceAddfOp/GoogleBenchmarkMain.cpp). |
 | Reduce Maxf | `ninja dl-op-reduce-maxf-benchmark` | This benchmark evaluates optimization strategies for the `reduce.maxf` operation. The benchmark size can be adjusted in [this file](./Ops/ReduceMaxfOp/GoogleBenchmarkMain.cpp). |
 | Softmax Exp Sum Div | `ninja dl-op-softmax-exp-sum-div-benchmark` | This benchmark evaluates optimization strategies for the `softmax.exp_sum_div` operation. The benchmark size can be adjusted in [this file](./Ops/SoftmaxExpSumDivOp/GoogleBenchmarkMain.cpp). |
+| RFFT | `ninja dl-op-linalg-RFFT-benchmark` | This benchmark evaluates the efficiency of the RFFT operator and verifies its correctness by comparing it with the RFFT function in numpy. The benchmark size can be adjusted in [this file](./Ops/RFFTOp/GoogleBenchmarkMain.cpp). And you can compare the time consumption with Numpy library in the `time_result.txt`.|
 
 ### Enter Python virtual environment
 We recommend you to use anaconda3 to create python virtual environment. You should install python packages as buddy-mlir/requirements.