Create Runtime after ConnectionHandlers

tests/test_priority_pool.py

@@ -8,7 +8,7 @@ from hivemind.moe.server.runtime import Runtime
 from petals.server.task_pool import PrioritizedTaskPool
 
 
-def _process_tasks(pools, results_valid):
+def _submit_tasks(pools, results_valid):
     torch.set_num_threads(1)
 
     futures = []
@@ -50,11 +50,13 @@ def test_priority_pools():
         PrioritizedTaskPool(dummy_pool_func, name="B", max_batch_size=1),
     )
 
+    # Simulate requests coming from ConnectionHandlers
+    proc = mp.context.ForkProcess(target=_submit_tasks, args=(pools, results_valid))
+    proc.start()
+
     runtime = Runtime({str(i): DummyBackend([pool]) for i, pool in enumerate(pools)}, prefetch_batches=0)
     runtime.start()
 
-    proc = mp.context.ForkProcess(target=_process_tasks, args=(pools, results_valid))
-    proc.start()
     proc.join()
     assert results_valid.is_set()