[Core] Async_output_proc: Add virtual engine support (towards pipeline parallel) (#7911)

vllm/core/scheduler.py

@@ -302,7 +302,7 @@ class Scheduler:
         cache_config: CacheConfig,
         lora_config: Optional[LoRAConfig],
         pipeline_parallel_size: int = 1,
-        output_proc_callback_fn: Optional[Callable] = None,
+        output_proc_callback: Optional[Callable] = None,
     ) -> None:
         self.scheduler_config = scheduler_config
         self.cache_config = cache_config
@@ -376,8 +376,8 @@ class Scheduler:
         # iterations. I.e. since the output processing is lagged one step,
         # we cannot reuse the cached objects immediately when the schedule()
         # is called again, but only when schedule() is called the second time.
-        self.output_proc_callback_fn = output_proc_callback_fn
-        self.use_async_output_proc = self.output_proc_callback_fn is not None
+        self.output_proc_callback = output_proc_callback
+        self.use_async_output_proc = self.output_proc_callback is not None
         self.num_cache_iters = 2 if self.use_async_output_proc else 1
 
         self.cache_id = 0
@@ -573,8 +573,8 @@ class Scheduler:
                     seq_group):
                 tmp = self.running
                 self.running = orig_running
-                assert self.output_proc_callback_fn is not None
-                self.output_proc_callback_fn(is_async=True)
+                assert self.output_proc_callback is not None
+                self.output_proc_callback()
                 self.running = tmp
 
             while not self._can_append_slots(seq_group):
@@ -1091,7 +1091,6 @@ class Scheduler:
         no_beam_search = seq_group.sampling_params is None or (
             seq_group.sampling_params.best_of == 1
             and not seq_group.sampling_params.use_beam_search)
-
         return no_beam_search
 
     def schedule(

vllm/engine/async_llm_engine.py

@@ -279,10 +279,16 @@ class _AsyncLLMEngine(LLMEngine):
         scheduler_outputs = cached_outputs.scheduler_outputs
         allow_async_output_proc = cached_outputs.allow_async_output_proc
 
+        ctx = self.scheduler_contexts[virtual_engine]
+
         # skip the scheduler if there are any remaining steps in the seq groups.
         # This ensures that the scheduler is only called again when the current
         # batch has completed.
         if not self._has_remaining_steps(seq_group_metadata_list):
+
+            # Clear outputs on scheduler iteration start
+            ctx.request_outputs.clear()
+
             (seq_group_metadata_list, scheduler_outputs,
              allow_async_output_proc
              ) = self.scheduler[virtual_engine].schedule()
@@ -290,8 +296,9 @@ class _AsyncLLMEngine(LLMEngine):
             # If current scheduler iteration has no async postprocessor,
             # then we need first to drain the pending async postprocessor
             # before moving forward
-            if not allow_async_output_proc and len(self.output_queue) > 0:
-                self._process_model_outputs(is_async=True)
+            if not allow_async_output_proc and len(ctx.output_queue) > 0:
+                self._process_model_outputs(virtual_engine=virtual_engine,
+                                            is_async=True)
 
             if (self.scheduler_config.is_multi_step
                     and scheduler_outputs.num_lookahead_slots > 0):
@@ -332,8 +339,8 @@ class _AsyncLLMEngine(LLMEngine):
                 last_sampled_token_ids=last_sampled_token_ids)
 
             if allow_async_output_proc:
-                execute_model_req.output_proc_callback_fn = \
-                    self._process_model_outputs
+                execute_model_req.async_callback = self.async_callback[
+                    virtual_engine]
 
             # Execute the model.
             output = await self.model_executor.execute_model_async(
@@ -343,9 +350,10 @@ class _AsyncLLMEngine(LLMEngine):
             if self.scheduler_config.is_multi_step:
                 self._update_cached_scheduler_output(virtual_engine, output)
         else:
-            if len(self.output_queue) > 0:
+            if len(ctx.output_queue) > 0:
                 assert not self.scheduler_config.is_multi_step
-                self._process_model_outputs(is_async=True)
+                self._process_model_outputs(virtual_engine=virtual_engine,
+                                            is_async=True)
             output = []
 
         # Finish the current step for all the sequence groups.
@@ -360,7 +368,7 @@ class _AsyncLLMEngine(LLMEngine):
                     virtual_engine] = SchedulerOutputState()
 
             # Cache results in engine
-            self.output_queue.append(
+            ctx.output_queue.append(
                 (output, seq_group_metadata_list, scheduler_outputs))
 
             if output and allow_async_output_proc:
@@ -372,7 +380,8 @@ class _AsyncLLMEngine(LLMEngine):
                     scheduler_outputs.scheduled_seq_groups)
 
             if not allow_async_output_proc:
-                self._process_model_outputs(is_async=False)
+                self._process_model_outputs(virtual_engine=virtual_engine,
+                                            is_async=False)
 
                 # Log stats.
                 self.do_log_stats(scheduler_outputs, output)
@@ -381,9 +390,17 @@ class _AsyncLLMEngine(LLMEngine):
                 self.do_tracing(scheduler_outputs)
 
         else:
-            self.request_outputs = []
+            ctx.request_outputs = []
+
+        if not self.has_unfinished_requests():
+            # Drain async postprocessor (if exists)
+            if len(ctx.output_queue) > 0:
+                assert not self.scheduler_config.is_multi_step
+                self._process_model_outputs(virtual_engine=virtual_engine,
+                                            is_async=True)
+            assert len(ctx.output_queue) == 0
 
-        return self.request_outputs
+        return ctx.request_outputs
 
     async def stop_remote_worker_execution_loop_async(self) -> None:
         """Stop the remote worker execution loop."""

vllm/engine/llm_engine.py

+import functools
 import time
 from collections import deque
 from contextlib import contextmanager
-from dataclasses import dataclass
+from dataclasses import dataclass, field
 from typing import (TYPE_CHECKING, Any, ClassVar, Deque, Dict, Iterable, List,
                     Mapping, Optional)
 from typing import Sequence as GenericSequence
@@ -88,6 +89,17 @@ class SchedulerOutputState:
     last_output: Optional[SamplerOutput] = None
 
 
+@dataclass
+class SchedulerContext:
+    output_queue: Deque[Tuple[List[SamplerOutput], List[SequenceGroupMetadata],
+                              SchedulerOutputs]] = field(
+                                  default_factory=lambda: deque())
+
+    request_outputs: List[Union[RequestOutput,
+                                EmbeddingRequestOutput]] = field(
+                                    default_factory=lambda: [])
+
+
 class LLMEngine:
     """An LLM engine that receives requests and generates texts.
 
@@ -350,9 +362,11 @@ class LLMEngine:
             Scheduler(
                 scheduler_config, cache_config, lora_config,
                 parallel_config.pipeline_parallel_size,
-                self._process_model_outputs
+                functools.partial(self._process_model_outputs,
+                                  virtual_engine=v_id,
+                                  is_async=True)
                 if model_config.use_async_output_proc else None)
-            for _ in range(parallel_config.pipeline_parallel_size)
+            for v_id in range(parallel_config.pipeline_parallel_size)
         ]
 
         # Metric Logging.
@@ -406,12 +420,17 @@ class LLMEngine:
             for _ in range(self.parallel_config.pipeline_parallel_size)
         ]
 
-        # Async output processing pointers
-        self.output_queue: Deque[Tuple[List[SamplerOutput],
-                                       List[SequenceGroupMetadata],
-                                       SchedulerOutputs]] = deque()
-        self.request_outputs: List[Union[RequestOutput,
-                                         EmbeddingRequestOutput]] = []
+        self.scheduler_contexts = [
+            SchedulerContext()
+            for _ in range(self.parallel_config.pipeline_parallel_size)
+        ]
+
+        self.async_callback = [
+            functools.partial(self._process_model_outputs,
+                              virtual_engine=v_id,
+                              is_async=True)
+            for v_id in range(self.parallel_config.pipeline_parallel_size)
+        ]
 
     def _initialize_kv_caches(self) -> None:
         """Initialize the KV cache in the worker(s).
@@ -1221,32 +1240,28 @@ class LLMEngine:
 
         return
 
-    def _process_model_outputs(self,
-                               is_async: bool,
-                               clear_outputs: bool = True) -> None:
+    def _process_model_outputs(self, virtual_engine: int,
+                               is_async: bool) -> None:
         """Apply the model output to the sequences in the scheduled seq groups.
 
+        virtual_engine: The engine id to operate on
         is_async: Indicates whether this postprocessor runs in 
             parallel with the GPU forward pass and is processing 
             tokens from the previous step. If this is true, then
             no tokens need to be appended since it is already done
             externally (before the next schedule() call)
-        clear_outputs: Sometimes existing outputs need to be combined 
-            with outputs of this call. This happens for postprocessor
-            draining at the final stage (like when sequences are finished)
         
         Returns RequestOutputs that can be returned to the client.
         """
         now = time.time()
 
-        if clear_outputs:
-            self.request_outputs.clear()
+        ctx: SchedulerContext = self.scheduler_contexts[virtual_engine]
 
-        if len(self.output_queue) == 0:
+        if len(ctx.output_queue) == 0:
             return None
 
         (outputs, seq_group_metadata_list,
-         scheduler_outputs) = self.output_queue.popleft()
+         scheduler_outputs) = ctx.output_queue.popleft()
 
         # Sanity check
         assert len(seq_group_metadata_list) == len(
@@ -1321,11 +1336,11 @@ class LLMEngine:
             if (seq_group.is_finished()
                     if self.step_return_finished_only else True):
                 request_output = RequestOutputFactory.create(seq_group)
-                self.request_outputs.append(request_output)
+                ctx.request_outputs.append(request_output)
 
         for seq_group in scheduler_outputs.ignored_seq_groups:
             request_output = RequestOutputFactory.create(seq_group)
-            self.request_outputs.append(request_output)
+            ctx.request_outputs.append(request_output)
 
         if is_async:
             # Log stats.
@@ -1421,29 +1436,43 @@ class LLMEngine:
                 "Pipeline parallelism is only supported through AsyncLLMEngine "
                 "as performance will be severely degraded otherwise.")
 
+        # For llm_engine, there is no pipeline parallel support, so the engine
+        # used is always 0
+        virtual_engine = 0
+
         # These are cached outputs from previous iterations. None if on first
         # iteration
-        cached_outputs = self.cached_scheduler_outputs[0]
+        cached_outputs = self.cached_scheduler_outputs[virtual_engine]
         seq_group_metadata_list = cached_outputs.seq_group_metadata_list
         scheduler_outputs = cached_outputs.scheduler_outputs
         allow_async_output_proc = cached_outputs.allow_async_output_proc
 
+        ctx = self.scheduler_contexts[virtual_engine]
+
         # Skip the scheduler if there are any remaining steps in the seq groups.
         # This ensures that the scheduler is only called again when the current
         # batch has completed.
         if not self._has_remaining_steps(seq_group_metadata_list):
+
+            # Clear outputs on scheduler iteration start
+            ctx.request_outputs.clear()
+
+            # Schedule iteration
             (seq_group_metadata_list, scheduler_outputs,
-             allow_async_output_proc) = self.scheduler[0].schedule()
+             allow_async_output_proc
+             ) = self.scheduler[virtual_engine].schedule()
 
-            if not allow_async_output_proc and len(self.output_queue) > 0:
-                self._process_model_outputs(is_async=True)
+            # Maybe switch from async mode to sync mode
+            if not allow_async_output_proc and len(ctx.output_queue) > 0:
+                self._process_model_outputs(virtual_engine=virtual_engine,
+                                            is_async=True)
 
             if (self.scheduler_config.is_multi_step
                     and scheduler_outputs.num_lookahead_slots > 0):
                 # cache the scheduler outputs for the next iteration if we have
                 # lookahead slots
                 self._cache_scheduler_outputs_for_multi_step(
-                    0, seq_group_metadata_list, scheduler_outputs,
+                    virtual_engine, seq_group_metadata_list, scheduler_outputs,
                     allow_async_output_proc)
 
         assert seq_group_metadata_list is not None
@@ -1454,14 +1483,14 @@ class LLMEngine:
 
         if not scheduler_outputs.is_empty():
             finished_requests_ids = self.scheduler[
-                0].get_and_reset_finished_requests_ids()
+                virtual_engine].get_and_reset_finished_requests_ids()
 
             # Check if we have a cached last_output from the previous iteration.
             # For supporting PP this is probably the best way to pass the
             # sampled_token_ids, as a separate broadcast over all the PP stages
             # will cause one virtual engine's microbatch to block the pipeline.
             last_sampled_token_ids = \
-                self._get_last_sampled_token_ids(0)
+                self._get_last_sampled_token_ids(virtual_engine)
 
             execute_model_req = ExecuteModelRequest(
                 seq_group_metadata_list=seq_group_metadata_list,
@@ -1476,20 +1505,24 @@ class LLMEngine:
                 last_sampled_token_ids=last_sampled_token_ids)
 
             if allow_async_output_proc:
-                execute_model_req.output_proc_callback_fn = \
-                    self._process_model_outputs
+                execute_model_req.async_callback = self.async_callback[
+                    virtual_engine]
 
             output = self.model_executor.execute_model(
                 execute_model_req=execute_model_req)
 
-            # we need to do this here so that last step's sampled_token_ids can
+            # We need to do this here so that last step's sampled_token_ids can
             # be passed to the next iteration for PP.
             if self.scheduler_config.is_multi_step:
-                self._update_cached_scheduler_output(0, output)
+                self._update_cached_scheduler_output(virtual_engine, output)
         else:
-            if len(self.output_queue) > 0:
+            # Nothing scheduled => If there is pending async postprocessor,
+            # then finish it here.
+            if len(ctx.output_queue) > 0:
                 assert not self.scheduler_config.is_multi_step
-                self._process_model_outputs(is_async=True)
+                self._process_model_outputs(virtual_engine=virtual_engine,
+                                            is_async=True)
+            # No outputs in this case
             output = []
 
         # Finish the current step for all the sequence groups.
@@ -1504,7 +1537,7 @@ class LLMEngine:
 
             # Add results to the output_queue
             # (for async or non-async postprocessing)
-            self.output_queue.append(
+            ctx.output_queue.append(
                 (output, seq_group_metadata_list, scheduler_outputs))
 
             if output and allow_async_output_proc:
@@ -1515,8 +1548,10 @@ class LLMEngine:
                     output[0], seq_group_metadata_list,
                     scheduler_outputs.scheduled_seq_groups)
 
+            # Check if need to run the usual non-async path
             if not allow_async_output_proc:
-                self._process_model_outputs(is_async=False)
+                self._process_model_outputs(virtual_engine=virtual_engine,
+                                            is_async=False)
 
                 # Log stats.
                 self.do_log_stats(scheduler_outputs, output)
@@ -1524,14 +1559,16 @@ class LLMEngine:
                 # Tracing
                 self.do_tracing(scheduler_outputs)
         else:
-            self.request_outputs = []
+            # Multi-step case
+            ctx.request_outputs = []
 
         if not self.has_unfinished_requests():
-            # Drain async postprocessor
-            if len(self.output_queue) > 0:
+            # Drain async postprocessor (if exists)
+            if len(ctx.output_queue) > 0:
                 assert not self.scheduler_config.is_multi_step
-                self._process_model_outputs(is_async=True, clear_outputs=False)
-            assert len(self.output_queue) == 0
+                self._process_model_outputs(virtual_engine=virtual_engine,
+                                            is_async=True)
+            assert len(ctx.output_queue) == 0
 
             # Stop the execute model loop in parallel workers until there are
             # more requests to process. This avoids waiting indefinitely in
@@ -1540,7 +1577,7 @@ class LLMEngine:
             # queued control plane messages, such as add/remove lora adapters.
             self.model_executor.stop_remote_worker_execution_loop()
 
-        return self.request_outputs
+        return ctx.request_outputs
 
     def _has_remaining_steps(
         self, seq_group_metadata_list: Optional[List[SequenceGroupMetadata]]

vllm/sequence.py

@@ -811,6 +811,9 @@ class SequenceGroup:
         self.is_single_seq = len(self.seqs) == 1
 
     def is_finished(self) -> bool:
+        if self.is_single_seq:
+            return self.seqs[0].is_finished()
+
         return all(seq.is_finished() for seq in self.seqs)
 
     def is_prefill(self) -> bool:
@@ -1290,8 +1293,8 @@ class ExecuteModelRequest(
     finished_requests_ids: List[str] = msgspec.field(default_factory=list)
     # The last sampled token ids for multi step decoding.
     last_sampled_token_ids: Optional[torch.Tensor] = None
-    # Async postprocessor
-    output_proc_callback_fn: Optional[Callable] = None
+    # Async callback
+    async_callback: Optional[Callable] = None
 
     @property
     def is_first_multi_step(self) -> bool:
@@ -1338,4 +1341,4 @@ class ExecuteModelRequest(
             finished_requests_ids=self.finished_requests_ids,
             last_sampled_token_ids=self.last_sampled_token_ids.clone()
             if self.last_sampled_token_ids is not None else None,
-            output_proc_callback_fn=self.output_proc_callback_fn)
+            async_callback=self.async_callback)

vllm/worker/model_runner.py

@@ -91,7 +91,7 @@ class ModelInputForGPU(ModelRunnerInputBase):
     request_ids_to_seq_ids: Optional[Dict[str, List[int]]] = None
     finished_requests_ids: Optional[List[str]] = None
     virtual_engine: int = 0
-    output_proc_callback_fn: Optional[Callable] = None
+    async_callback: Optional[Callable] = None
 
     def as_broadcastable_tensor_dict(self) -> Dict[str, Any]:
         tensor_dict = {
@@ -1457,8 +1457,8 @@ class ModelRunner(GPUModelRunnerBase[ModelInputForGPUWithSamplingMetadata]):
         if not self.is_driver_worker:
             return []
 
-        if model_input.output_proc_callback_fn is not None:
-            model_input.output_proc_callback_fn(is_async=True)
+        if model_input.async_callback is not None:
+            model_input.async_callback()
 
         # Sample the next token.
         output: SamplerOutput = self.model.sample(

vllm/worker/worker_base.py

@@ -263,11 +263,10 @@ class LocalOrDistributedWorkerBase(WorkerBase):
             broadcast_data.update(kwargs)
             broadcast_tensor_dict(broadcast_data, src=0)
 
-        if execute_model_req.output_proc_callback_fn:
+        if execute_model_req.async_callback:
             model_input = dataclasses.replace(  # type: ignore
                 model_input,
-                output_proc_callback_fn=execute_model_req.
-                output_proc_callback_fn)
+                async_callback=execute_model_req.async_callback)
 
         return model_input, worker_input, kwargs