Add "benchmark" to time atomic operations

The intent here is to confirm that the compiler behaves as expected,
and to guide optimization efforts, particularly where there is a
choice between primitives.

Test: Built and ran benchmark on Angler repeatedly. Manually confirmed
that the compiler behaves roughly as expected.

Change-Id: I059b245d1ba8296e9b28602559b53eafafe0a30f
diff --git a/benchmarks/Android.bp b/benchmarks/Android.bp
index 12f4940..3f95aa1 100644
--- a/benchmarks/Android.bp
+++ b/benchmarks/Android.bp
@@ -25,6 +25,7 @@
         "-Wunused",
     ],
     srcs: [
+        "atomic_benchmark.cpp",
         "math_benchmark.cpp",
         "property_benchmark.cpp",
         "pthread_benchmark.cpp",