test to check listen-accept model

tests/test-server-listen-accept.py

@@ -0,0 +1,73 @@
+# Copyright (c) 2018, NVIDIA CORPORATION. All rights reserved.
+# See file LICENSE for terms.
+
+import call_myucp as ucp
+import time
+import argparse
+
+def notify_completion(what_completed):
+    print(what_completed)
+
+def send_recv(msg_log, is_server, is_cuda):
+    buffer_region = ucp.buffer_region()
+    if is_cuda:
+        buffer_region.alloc_cuda(1 << msg_log)
+    else:
+        buffer_region.alloc_host(1 << msg_log)
+    msg = ucp.ucp_msg(buffer_region)
+    if 1 == is_server:
+        msg.set_mem(0, 1 << msg_log)
+        msg.send(1 << msg_log)
+    else:
+        msg.set_mem(1, 1 << msg_log)
+        msg.recv(1 << msg_log)
+    msg.wait()
+
+    if is_cuda:
+        buffer_region.free_cuda()
+    else:
+        buffer_region.free_host()
+
+max_msg_log = 23
+
+parser = argparse.ArgumentParser()
+parser.add_argument('-s','--server', help='enter server ip', required=False)
+parser.add_argument('-p','--port', help='enter server port number', required=False)
+args = parser.parse_args()
+
+## initiate ucp
+init_str = ""
+is_server = 0
+if args.server is None:
+    is_server = 1
+else:
+    is_server = 0
+    init_str = args.server
+
+## setup endpoints
+ucp.init(init_str.encode(), is_server, server_listens = 1)
+
+if 0 == is_server:
+    #connect to server
+    ucp.get_endpoint(init_str.encode(), int(args.port))
+    is_cuda = False
+    send_recv(max_msg_log, is_server, is_cuda)
+    is_cuda = True
+    send_recv(max_msg_log, is_server, is_cuda)
+else:
+    #setup callback
+    ucp.wait_for_client()
+    is_cuda = False
+    send_recv(max_msg_log, is_server, is_cuda)
+    is_cuda = True
+    send_recv(max_msg_log, is_server, is_cuda)
+
+#ucp.set_callback(notify_completion)
+
+ucp.destroy_ep()
+ucp.fin()
+
+if args.server is None:
+    print("Server Finalized")
+else:
+    print("Client Finalized")