diff --git a/acceptance/router_benchmark/test.py b/acceptance/router_benchmark/test.py
index dbea3285f8..dec5f05e75 100755
--- a/acceptance/router_benchmark/test.py
+++ b/acceptance/router_benchmark/test.py
@@ -63,6 +63,27 @@ def _run(self):
         # Give some time for the topology to start.
         self.await_connectivity()
 
+        # Fetch and log the number of cores used by Go. This may inform performance
+        # modeling later.
+        promQuery = urlencode({
+            'query': 'avg by (job) (go_sched_maxprocs_threads{job="BR"})'
+        })
+
+        conn = HTTPConnection("localhost:9090")
+        conn.request('GET', f'/api/v1/query?{promQuery}')
+        resp = conn.getresponse()
+        if resp.status != 200:
+            raise RuntimeError(f'Unexpected response: {resp.status} {resp.reason}')
+
+        pld = json.loads(resp.read().decode('utf-8'))
+        results = pld['data']['result']
+        numCores = 0
+        for result in results:
+            _, val = result['value']
+            numCores = int(float(val))
+
+        logger.info(f'Router Cores: {numCores}')
+
         # Start as-transiting load. With the router_bm topology
 
         # The subset noncore#nonlocalcore gives us outgoing traffic at each