[tools] Handle benchmark runs reporting more than one result.

tools/perf-to-html.py

Index: tools/perf-to-html.py
diff --git a/tools/perf-to-html.py b/tools/perf-to-html.py
index 63faeb1d6602e232cd626f1aa678ec53a94b93e7..7ec9c50f218fe123b37fd254c96731c23bdd5483 100755
--- a/tools/perf-to-html.py
+++ b/tools/perf-to-html.py
@@ -115,8 +115,8 @@ class Benchmark:
     self.name_ = name
     self.tests_ = {}
     for test in data:
-      # strip off "<name>/" prefix
-      test_name = test.split("/")[1]
+      # strip off "<name>/" prefix, allowing for subsequent "/"s
+      test_name = test.split("/", 1)[1]
       self.appendResult(test_name, data[test])
 
   # tests is a dictionary of Results