stdlib: Move _m5.stats.processDumpQueue to call-once

This commit addresses Jason's comment
(https://github.com/gem5/gem5/pull/996#discussion_r1613870880) which
highlighted putting the `_m5.stats.processDumpQueue` call in the
iteration through the `root` object in `get_simstat` caused this
function be potentially called many times when it only needs to be
called once. This chance moved this call to just before this iteration
and will tehrefore only be called once (if required) per `get_simstat`
execution.

Change-Id: I16908b6dee063a0df7877a19e215883963bfb081

src/python/m5/stats/gem5stats.py

@@ -373,6 +373,10 @@ def get_simstat(
     :Returns: The SimStat Object of the current simulation.
 
     """
+
+    if prepare_stats:
+        _m5.stats.processDumpQueue()
+
     stats_map = {}
     for r in root:
         creation_time = datetime.now()
@@ -384,9 +388,6 @@ def get_simstat(
         simulated_begin_time = int(final_tick - sim_ticks)
         simulated_end_time = int(final_tick)
 
-        if prepare_stats:
-            _m5.stats.processDumpQueue()
-
         if prepare_stats:
             if isinstance(r, list):
                 for obj in r: