Remove ranks-cpus warning as can be inaccurate (#66)

The warning is misleading and causes users to run simulation with more
resources than needed. Since the magic cutoff is neither a lower bound,
nor an upper bound (by many orders of magnitude), we should simply not
print anything.

For cases where resource consumption is hard to estimate we encourage
user to try the new dry-run

neurodamus/node.py

@@ -369,13 +369,6 @@ def compute_load_balance(self):
         # Info about the cells to be distributed
         target_spec = TargetSpec(circuit.CircuitTarget)
         target = self.target_manager.get_target(target_spec)
-        cell_count = None
-
-        if not target.is_void():
-            cell_count = target.gid_count()
-            if cell_count > 100 * MPI.size:
-                logging.warning("Your simulation has a very high count of cells per CPU. "
-                                "Please consider launching it in a larger MPI cluster")
 
         # Check / set load balance mode
         lb_mode = LoadBalance.select_lb_mode(SimConfig, self._run_conf, target)