another race

Author: 3manifold

src/cpu/primitives.cc

@@ -294,6 +294,16 @@ namespace ctranslate2 {
   template<>
   template<>
   void primitives<Device::CPU>::relu(const float* x, float* y, dim_t size) {
+
+    int sum = 0;
+    int n = 10000;
+
+#pragma omp parallel for reduction(+:sum)
+    for (int i = 0; i < n; i++) {
+      sum += i;  // Each thread has private copy, combined at end
+    }
+
+
       float sum = 0.0f;  // Shared state
 
     cpu::parallel_for(0, size, cpu::GRAIN_SIZE,