Roll benchmark to fix compile warnings in clang-13 (#7202)

Fixes https://github.com/google/benchmark/issues/1172

Also update to using state.threads()
diff --git a/SUBMODULE_VERSIONS.txt b/SUBMODULE_VERSIONS.txt
index 5e47924..bdf58da 100644
--- a/SUBMODULE_VERSIONS.txt
+++ b/SUBMODULE_VERSIONS.txt
@@ -1,4 +1,4 @@
-39c8d58a76ce7c3398786ee96a452668e8392af7 third_party/benchmark
+0baacde3618ca617da95375e0af13ce1baadea47 third_party/benchmark
 5916273f79a21551890fd3d56fc5375a78d1598d third_party/cpuinfo
 62a7f88611af08c68be5485ce6d9c7e92474eab8 third_party/flatcc
 aa533abfd4232b01f9e57041d70114d5a77e6de0 third_party/googletest
diff --git a/iree/base/internal/synchronization_benchmark.cc b/iree/base/internal/synchronization_benchmark.cc
index f2674da..9bdc13a 100644
--- a/iree/base/internal/synchronization_benchmark.cc
+++ b/iree/base/internal/synchronization_benchmark.cc
@@ -135,7 +135,7 @@
     // To achieve this amount of local work is multiplied by number of threads
     // to keep ratio between local work and critical section approximately
     // equal regardless of number of threads.
-    SpinDelay(100 * state.threads, &local);
+    SpinDelay(100 * state.threads(), &local);
     RaiiLocker<MutexType> locker(&mu);
     SpinDelay(static_cast<int>(state.range(0)), &data);
   }
@@ -179,7 +179,7 @@
     // To achieve this amount of local work is multiplied by number of threads
     // to keep ratio between local work and critical section approximately
     // equal regardless of number of threads.
-    SpinDelay(100 * state.threads, &local);
+    SpinDelay(100 * state.threads(), &local);
     RaiiLocker<MutexType> locker(&shared->mu);
     SpinDelay(static_cast<int>(state.range(0)), &shared->data);
   }
diff --git a/third_party/benchmark b/third_party/benchmark
index 39c8d58..0baacde 160000
--- a/third_party/benchmark
+++ b/third_party/benchmark
@@ -1 +1 @@
-Subproject commit 39c8d58a76ce7c3398786ee96a452668e8392af7
+Subproject commit 0baacde3618ca617da95375e0af13ce1baadea47