diff --git a/xtuner/v1/engine/train_engine.py b/xtuner/v1/engine/train_engine.py
index 184ee35db..9032332b8 100644
--- a/xtuner/v1/engine/train_engine.py
+++ b/xtuner/v1/engine/train_engine.py
@@ -318,11 +318,12 @@ def async_save_hf(
         save_dtype: torch.dtype = torch.bfloat16,
         cleanup_hf_dirs: Sequence[str | Path] = (),
     ) -> AsyncHFSaveHandle:
-        return self.model.async_save_hf(
-            hf_dir=hf_dir,
-            save_dtype=save_dtype,
-            cleanup_hf_dirs=cleanup_hf_dirs,
-        )
+        with profile_time_and_memory(f"[Async saving HF to {hf_dir} launch cost]"):
+            return self.model.async_save_hf(
+                hf_dir=hf_dir,
+                save_dtype=save_dtype,
+                cleanup_hf_dirs=cleanup_hf_dirs,
+            )
 
     def wait_async_hf(self, handle: AsyncHFSaveHandle | None = None) -> Path | None:
         return self.model.wait_async_hf(handle)
diff --git a/xtuner/v1/model/base.py b/xtuner/v1/model/base.py
index cad387a99..336ebfebf 100644
--- a/xtuner/v1/model/base.py
+++ b/xtuner/v1/model/base.py
@@ -762,6 +762,7 @@ def _run_async_hf_writer(
         cleanup_done_path: Path,
         rank: int,
     ) -> None:
+        log_rank0.info(f"[Async saving HF to {tmp_hf_dir} writer] started")
         try:
             set_async_save_process_qos()
             self._cleanup_async_hf_dirs_before_write(
@@ -775,7 +776,9 @@ def _run_async_hf_writer(
                 weight_map=weight_map,
                 status_path=status_path,
             )
+            log_rank0.info(f"[Async saving HF to {tmp_hf_dir} writer] finished")
         except Exception as exc:
+            log_rank0.error(f"[Async saving HF to {tmp_hf_dir} writer] failed: {exc}")
             status = {"rank": rank, "ok": False, "error": str(exc), "weight_map": {}}
             with status_path.open("w") as f:
                 f.write(json.dumps(status, indent=2))
diff --git a/xtuner/v1/model/compose/base.py b/xtuner/v1/model/compose/base.py
index 68c9957ba..b970aaad5 100644
--- a/xtuner/v1/model/compose/base.py
+++ b/xtuner/v1/model/compose/base.py
@@ -258,6 +258,7 @@ def _run_async_hf_compose_writer(
         cleanup_done_path: Path,
         rank: int,
     ) -> None:
+        log_rank0.info(f"[Async saving HF to {tmp_hf_dir} writer] started")
         try:
             set_async_save_process_qos()
             self._cleanup_async_hf_dirs_before_write(
@@ -270,7 +271,9 @@ def _run_async_hf_compose_writer(
             status = {"rank": rank, "ok": True, "error": "", "weight_map": merged_weight_map}
             with status_path.open("w") as f:
                 f.write(json.dumps(status, indent=2))
+            log_rank0.info(f"[Async saving HF to {tmp_hf_dir} writer] finished")
         except Exception as exc:
+            log_rank0.error(f"[Async saving HF to {tmp_hf_dir} writer] failed: {exc}")
             status = {"rank": rank, "ok": False, "error": str(exc), "weight_map": {}}
             with status_path.open("w") as f:
                 f.write(json.dumps(status, indent=2))