Remove the "sleep" delay in the oob connection procedure. This shouldn't cause any problems, especially for launches of less than 1000 processes.

Please report any abnormal behavior during launch, though, as we would like to understand what (if any) impact is seen. I couldn't see any on small jobs (the modulo functions render this number down pretty low).

This commit was SVN r10763.

orte/mca/oob/tcp/oob_tcp.c

@@ -753,13 +753,15 @@ int mca_oob_tcp_init(void)
     char *tmp, *tmp2, *tmp3;
     size_t i, num_tokens;
 
+    /** turn this off for scalability testing */
+#if 0
     /* random delay to stagger connections back to seed */
 #if defined(__WINDOWS__)
     sleep((orte_process_info.my_name->vpid % orte_process_info.num_procs % 1000) * 1000);
 #else
     usleep((orte_process_info.my_name->vpid % orte_process_info.num_procs % 1000) * 1000);
 #endif
-
+#endif
     /* get my jobid */
     if (ORTE_SUCCESS != (rc = orte_ns.get_jobid(&jobid,
                                                 orte_process_info.my_name))) {