chore: Improve the AutoTuner log information.

* Change fallback alert from DEBUG level to WARNING level  and only do once.
* Add debug information for profiling cache right after warmup phase.
* Change the level of exception message during tactic profiling from ERROR to WARNING. And all exception details are pushed to DEBUG.
* Other trivial refinements.

Signed-off-by: Yukun He <[email protected]>

Author: hyukn

tensorrt_llm/_torch/autotuner.py

@@ -197,13 +197,13 @@ def autotune(tune_mode: bool = True):
     AutoTuner.get().is_tuning_mode = tune_mode
     autotune_enabled = tune_mode and not old_mode
     if autotune_enabled:
-        logger.info("[Autotuner]: Autotuning process starts ...")
+        logger.info("[Autotuner] Autotuning process starts ...")
     try:
         yield
     finally:
         AutoTuner.get().is_tuning_mode = old_mode
         if autotune_enabled:
-            logger.info("[Autotuner]: Autotuning process ends")
+            logger.info("[Autotuner] Autotuning process ends")
 
 
 @dataclass
@@ -350,16 +350,11 @@ def choose_one(self, custom_op: str, runners: List[TunableRunner],
             runner = runners[runner_id]
             # TODO: check the stored runner and tactic can implement this shape here
             # Should not directly try (runner, tactic) here, or it will hurt a lot of inference perf.
-
-            # Record the cache miss config.
-            # Expect no cache miss in inference. Thus, any cache miss should be recorded.
-            if not is_cache_hit:
-                logger.debug(
-                    f"[AutoTunner]: Using fallback tactic for {custom_op} with input shapes {input_shapes}"
-                )
-                logger.debug(
-                    f"[AutoTunner]: Generated key{AutoTuner._get_cache_key(custom_op, runners[0], input_shapes, tuning_config)}"
-                )
+            if not is_cache_hit and len(self.profiling_cache) > 0:
+                # Only log once for each custom op and only when cache is not empty
+                logger.warning_once(
+                    f"[AutoTunner] Using the fallback tactic, due to cache miss on input shapes={input_shapes}",
+                    key=(custom_op))
             return runner, tactic
 
         assert len(runners) > 0, "At least one runner is required"
@@ -370,6 +365,8 @@ def choose_one(self, custom_op: str, runners: List[TunableRunner],
         # Record the total configs to try
         self.stats.tuned_op_total_configs[custom_op] = len(profiles)
 
+        new_tuning_failuer_occured = False
+        
         for p in profiles:
             tensors = self._prepare_input_tensors(p, inputs)
             is_cache_hit, runner_id, tactic, _ = self.search_cache(
@@ -396,11 +393,13 @@ def choose_one(self, custom_op: str, runners: List[TunableRunner],
                         except Exception as e:
                             shapes = self._get_input_sizes(tensors)
 
-                            logger.error(
-                                f"[Autotuner]: Failed when profiling {r} {tac}, shapes={shapes}. Error occurred: {e}"
+                            logger.warning(
+                                f"[Autotuner] Failed when profiling runner={r}, tactic={tac}, shapes={shapes}. Set TLLM_LOG_LEVEL=DEBUG for more details."
                             )
+                            logger.debug(f"[Autotuner] Exception captured: {e}")
 
                             # Record the failed profiling combinations
+                            new_tuning_failuer_occured = True
                             if custom_op not in self.stats.failed_profiling_count:
                                 self.stats.failed_profiling_count[
                                     custom_op] = set()
@@ -426,8 +425,24 @@ def choose_one(self, custom_op: str, runners: List[TunableRunner],
                         custom_op] = self.stats.tuned_op_successful_configs.get(
                             custom_op, 0) + 1
                     logger.debug(
-                        f"[Autotuner]: profiling chosen runner: {runners[runner_id]} {tactic} for {cache_key}"
+                        f"[Autotuner] Profiling runner={runners[runner_id]}, tactic={tactic} for cache_key={cache_key}."
                     )
+                else:
+                    logger.warning(
+                        f"[Autotuner] No valid runner/tactic was found for custom_op={custom_op}, input_shapes={input_shapes}. "
+                        f"At least one valid (runner, tactic) pair is required. "
+                        f"If get_valid_tactics is intended to return empty list, please ensure that this profile is not valid for the custom_op "
+                        f"and should not occurs during the inference stage, or fallback tactic is implemented. Otherwise, the the tuning process will crash."
+                    )
+
+        # If failed profiling tactics occurs, log the error.
+        if new_tuning_failuer_occured:
+            logger.error(
+                f"[Autotuner] New tuning error occurs: Total failed profiling tactics occurs: {self.stats.failed_profiling_count[custom_op]} for custom_op={custom_op}. "
+                f"This will not block the tuning process. "
+                f"Please set TLLM_LOG_LEVEL=WARNING to find out when the tactic profiling fails. "
+                f"Set TLLM_LOG_LEVEL=DEBUG to get more details of the failures."
+            )
 
         # Get the best runner and tactic from cache
         # If no valid tactic is found, the fallback runner and tactic will be used
@@ -487,7 +502,7 @@ def _profile_single_kernel(self, runner: TunableRunner,
 
         shapes = self._get_input_sizes(inputs)
         logger.debug(
-            f"[Autotuner]: profiling {runner} {tactic}, shapes={shapes}, avg_time {avg_time}"
+            f"[Autotuner] Profiled runner={runner}, tactic={tactic}, shapes={shapes}: {avg_time}ms."
         )
 
         return avg_time
@@ -557,7 +572,7 @@ def _optimization_profiles(
                 p.shapes[spec.input_idx][spec.dim_idx] = DynamicDim(
                     min_value, opt_value, max_value)
             generated_profiles.append(p)
-            logger.debug(f"[Autotuner]: generated profile: {p}")
+            logger.debug(f"[Autotuner] Generated profile: {p}")
         return generated_profiles
 
     @classmethod
@@ -649,3 +664,12 @@ def clear_cache(self) -> None:
     def reset_statistics(self) -> None:
         """Reset all statistics counters."""
         self.stats = AutoTunerStatistics()
+
+    def print_profiling_cache(self):
+        logger.debug(f"[Autotuner] The profiling_cache entries:")
+        logger.debug(
+            f"[Autotuner] Cache contents: (custom_op, runner, hash(attributes), shape_profiles) -> (runner_id, tactic, shape_profile(ignored))"
+        )
+        for key, value in self.profiling_cache.items():
+            runner_id, tactic, _ = value
+            logger.debug(f"[Autotuner] {key}: ({runner_id}, {tactic})")

tensorrt_llm/_torch/pyexecutor/model_engine.py

@@ -693,15 +693,16 @@ def disable_optimization(backend: Backend):
                             # No KV cache space!
                             pass
                         else:
-                            logger.info(
-                                f"Run autotuning warmup for batch size={1}")
                             self.forward(batch,
                                          new_tensors_device=None,
                                          resource_manager=resource_manager)
                             torch.cuda.synchronize()
 
-                    logger.info(f"Autotuner Cache size after warmup " +
-                                str(len(AutoTuner.get().profiling_cache)))
+                    logger.info(
+                        f"[Autotuner] Cache size after warmup is {len(AutoTuner.get().profiling_cache)}"
+                    )
+
+                AutoTuner.get().print_profiling_cache()
 
             if not (self._run_cuda_graphs
                     or self._torch_compile_piecewise_cuda_graph):