chore(ci): use name field instead of run_name in benchmark results

In case of benchmark_repetitions > 1 using 'run_name' will return
always the same name for a given benchmark case. It won't make the
distinction between the iterations and the aggregated values like
mean, median or stdev. Using 'name' field fix this behavior.

ci/benchmark_parser.py

@@ -40,7 +40,7 @@ def parse_results(raw_results):
     """
     raw_results = json.loads(raw_results.read_text())
     return [
-        {"value": res["cpu_time"], "test": res["run_name"]}
+        {"value": res["cpu_time"], "test": res["name"]}
         for res in raw_results["benchmarks"]
     ]