diff --git a/string/test/test.cpp b/string/test/test.cpp
new file mode 100644
index 0000000000000000000000000000000000000000..ec797ef42e468cb310141b6d6879791f9ee478e3
--- /dev/null
+++ b/string/test/test.cpp
@@ -0,0 +1,84 @@
+// 编译测试代码的时候要加O3优化，优先使用preload方法覆盖系统memcpy。
+#include <string.h>
+#include <cstdio>
+#include <cstdlib>
+#include <time.h>
+#include <sys/mman.h>
+
+#define NUM_TESTS 1000000
+
+
+extern "C" void* memcpy1(void* dst, const void* src, size_t size);
+
+// 如果你要验证预取的收益，需要在执行memcpy之后清理缓存。
+void flush_cache() {
+    volatile char* flush = (volatile char*)malloc(1024 * 1024); // 1MB
+    for (int i = 0; i < 1024 * 1024; i += 64) { 
+        flush[i] = i;
+    }
+    free((void*)flush);
+}
+
+
+
+void test_memcpy_performance(const size_t sizes[], size_t num_sizes) {
+
+    struct timespec start, end;
+    long long total_time_ns;
+
+    for (size_t i = 0; i < num_sizes; ++i) {
+        size_t size = sizes[i];
+        char* src = (char *)malloc(size);
+        char* dst = (char *)malloc(size);
+        if (src == NULL || dst == NULL) {
+            fprintf(stderr, "Memory allocation failed\n");
+            exit(EXIT_FAILURE);
+        }
+        for (size_t j = 0; j < size; j++) {
+            src[j] = (char)(j % 256);
+        }
+
+        clock_gettime(CLOCK_MONOTONIC, &start);
+        for (int j = 0; j < NUM_TESTS; ++j) {
+            memcpy1(dst, src, size);
+            // flush_cache();
+        }
+        clock_gettime(CLOCK_MONOTONIC, &end);
+        total_time_ns = (end.tv_sec - start.tv_sec) * 1000000000LL + (end.tv_nsec - start.tv_nsec);
+        double avg_sys = (double)total_time_ns / NUM_TESTS;
+
+        printf("%.3f\n", size, avg_sys);
+
+        if (memcmp(src, dst, size) != 0) {
+            fprintf(stderr, "Error: memcpy verification failed for size %zu\n", size);
+            exit(EXIT_FAILURE);
+        }
+
+        free(src);
+        free(dst);
+    }
+}
+
+
+int main(int argc, char* argv[]) {
+
+
+
+      int start = 0;
+      int end = 200;
+      int size = end - size + 1;
+      size_t sizes[size + 10];
+      for (start = 0; start < size; start++)
+      {
+          sizes[start] = start + 1;  
+      }
+  
+      size_t additional_sizes[] = {256, 500, 512, 1024, 2000, 4096, 16384, 30000, 65536, 1024*1024};
+      int num_additional = sizeof(additional_sizes) / sizeof(additional_sizes[0]);
+      for (int i = 0; i < num_additional; i++) {
+         sizes[size + i] = additional_sizes[i];
+      }
+  
+      size_t total_sizes = size + num_additional;
+      test_memcpy_performance(sizes, total_sizes);
+}
\ No newline at end of file