Simplified the code a little by replacing some stuff with calls to

parallel_for_blocked()

dlib/svm/structural_svm_distributed.h

@@ -216,10 +216,6 @@ namespace dlib
 
                         data.num = problem.get_num_samples();
 
-                        // how many samples to process in a single task (aim for 4 jobs per worker)
-                        const long num_workers = std::max(1UL, tp.num_threads_in_pool());
-                        const long block_size = std::max(1L, data.num/(num_workers*4));
-
                         const uint64 start_time = ts.get_timestamp();
 
                         // pick fastest buffering strategy
@@ -233,11 +229,7 @@ namespace dlib
                         }
 
                         binder b(*this, req, data, buffer_subgradients_locally);
-                        for (long i = 0; i < data.num; i+=block_size)
-                        {
-                            tp.add_task(b, &binder::call_oracle, i, std::min(i + block_size, data.num));
-                        }
-                        tp.wait_for_all_tasks();
+                        parallel_for_blocked(tp, 0, data.num, b, &binder::call_oracle);
 
                         const uint64 stop_time = ts.get_timestamp();
                         if (buffer_subgradients_locally)

dlib/svm/structural_svm_problem_threaded.h

@@ -116,11 +116,6 @@ namespace dlib
         ) const
         {
             ++num_iterations_executed;
-            const long num = this->get_num_samples();
-
-            // how many samples to process in a single task (aim for 4 jobs per worker)
-            const long num_workers = std::max(1UL, tp.num_threads_in_pool());
-            const long block_size = std::max(1L, num/(num_workers*4));
 
             const uint64 start_time = ts.get_timestamp();
 
@@ -133,13 +128,8 @@ namespace dlib
                 buffer_subgradients_locally = !buffer_subgradients_locally;
             }
 
-
             binder b(*this, w, subgradient, total_loss, buffer_subgradients_locally);
-            for (long i = 0; i < num; i+=block_size)
-            {
-                tp.add_task(b, &binder::call_oracle, i, std::min(i+block_size, num));
-            }
-            tp.wait_for_all_tasks();
+            parallel_for_blocked(tp, 0, this->get_num_samples(), b, &binder::call_oracle);
 
             const uint64 stop_time = ts.get_timestamp();