diff --git a/examples/mpm_lagrangian_forces.py b/examples/mpm_lagrangian_forces.py
index d73b3deba8f3a..7c8a844d92441 100644
--- a/examples/mpm_lagrangian_forces.py
+++ b/examples/mpm_lagrangian_forces.py
@@ -1,7 +1,7 @@
 import taichi as ti
 import numpy as np
 
-ti.init(arch=ti.gpu, kernel_profiler=True)
+ti.init(arch=ti.gpu)
 
 dim = 2
 quality = 1  # Use a larger integral number for higher quality
diff --git a/python/taichi/lang/__init__.py b/python/taichi/lang/__init__.py
index 5f1b50efa7eda..8c6c35c76973b 100644
--- a/python/taichi/lang/__init__.py
+++ b/python/taichi/lang/__init__.py
@@ -32,7 +32,6 @@
                         'a.normalized()')(Matrix.normalized)
 
 cfg = default_cfg()
-current_cfg = current_cfg()
 x86_64 = core.x64
 x64 = core.x64
 arm64 = core.arm64
diff --git a/python/taichi/misc/util.py b/python/taichi/misc/util.py
index dc8bc1969af53..ac08ae4459250 100644
--- a/python/taichi/misc/util.py
+++ b/python/taichi/misc/util.py
@@ -278,6 +278,24 @@ def get_kernel_stats():
     return ti_core.get_kernel_stats()
 
 
+def print_async_stats(include_kernel_profiler=False):
+    import taichi as ti
+    if include_kernel_profiler:
+        ti.kernel_profiler_print()
+        print()
+    stat = ti.get_kernel_stats()
+    counters = stat.get_counters()
+    print('=======================')
+    print('Async benchmark metrics')
+    print('-----------------------')
+    print(f'Async mode:           {ti.current_cfg().async_mode}')
+    print(f'Kernel time:          {ti.kernel_profiler_total_time():.3f} s')
+    print(f'Tasks launched:       {int(counters["launched_tasks"])}')
+    print(f'Instructions emitted: {int(counters["codegen_statements"])}')
+    print(f'Tasks compiled:       {int(counters["codegen_offloaded_tasks"])}')
+    print('=======================')
+
+
 __all__ = [
     'vec',
     'veci',