diff --git a/src/UtilsCtrl/ThreadPool/UThreadPool.cpp b/src/UtilsCtrl/ThreadPool/UThreadPool.cpp
index f39acb6e..2eb3d016 100644
--- a/src/UtilsCtrl/ThreadPool/UThreadPool.cpp
+++ b/src/UtilsCtrl/ThreadPool/UThreadPool.cpp
@@ -209,16 +209,12 @@ CStatus UThreadPool::releaseSecondaryThread(CInt size) {
 CIndex UThreadPool::dispatch(const CIndex origIndex) {
     CIndex realIndex = 0;
     if (CGRAPH_DEFAULT_TASK_STRATEGY == origIndex) {
-        realIndex = cur_index_++;
+        realIndex = cur_index_.fetch_add(1, std::memory_order_relaxed) % config_.max_thread_size_;
         if (realIndex >= 0 && realIndex < config_.default_thread_size_
             && primary_threads_[realIndex]->is_running_) {
             // 如果是默认调度，并且被放置到 正在running 的pt中，则切换为 trigger_one 的策略，防止阻塞
             realIndex = CGRAPH_TRIGGER_ALL_THREAD_STRATEGY;
         }
-
-        if (cur_index_ >= config_.max_thread_size_ || cur_index_ < 0) {
-            cur_index_ = 0;
-        }
     } else {
         realIndex = origIndex;
     }
diff --git a/src/UtilsCtrl/ThreadPool/UThreadPool.h b/src/UtilsCtrl/ThreadPool/UThreadPool.h
index adca1708..800ca1e8 100644
--- a/src/UtilsCtrl/ThreadPool/UThreadPool.h
+++ b/src/UtilsCtrl/ThreadPool/UThreadPool.h
@@ -199,7 +199,7 @@ class UThreadPool : public UThreadObject {
 
 private:
     CBool is_init_ { false };                                                       // 是否初始化
-    CInt cur_index_ = 0;                                                            // 记录放入的线程数
+    std::atomic<CInt> cur_index_ { 0 };                                          // 记录放入的线程数
     UAtomicQueue<UTask> task_queue_;                                                // 用于存放普通任务
     UAtomicPriorityQueue<UTask> priority_task_queue_;                               // 运行时间较长的任务队列，仅在辅助线程中执行
     std::vector<UThreadPrimaryPtr> primary_threads_;                                // 记录所有的主线程
diff --git a/test/Performance/test-performance-04.cpp b/test/Performance/test-performance-04.cpp
index 049e93aa..677879ca 100644
--- a/test/Performance/test-performance-04.cpp
+++ b/test/Performance/test-performance-04.cpp
@@ -15,7 +15,7 @@ void test_performance_04() {
 
     const int layer = 8;
     const int nodePerLayer = 8;
-    const int runTimes = 100000;
+    const int runTimes = 300000;
 
     UThreadPoolConfig config;
     config.default_thread_size_ = nodePerLayer;
diff --git a/tutorial/T01-Simple.cpp b/tutorial/T01-Simple.cpp
index 006a2661..f3930350 100644
--- a/tutorial/T01-Simple.cpp
+++ b/tutorial/T01-Simple.cpp
@@ -42,7 +42,7 @@ void tutorial_simple() {
 
     /* 运行流图信息。初始化后，支持多次循环计算 */
     for (int i = 0; i < 3; i++) {
-        status += pipeline->run();
+        status = pipeline->run();
         CGRAPH_ECHO("==== tutorial_simple, loop : [%d], and run status = [%d].", i + 1, status.getCode());
     }