trigger profiling on abort

Summary:
record the profile trace if the training process receives SIGABRT e.g. when Process Group watchdog aborts the process

Author: tushar00jain

torchtitan/experiments/forge/example_train.py

@@ -282,12 +282,13 @@ def train(self):
         self.checkpointer.load(step=job_config.checkpoint.load_step)
         logger.info(f"Training starts at step {self.step + 1}.")
 
+        torch_profiler = maybe_enable_profiling(
+            job_config.profiling,
+            global_step=self.step,
+            base_folder=job_config.job.dump_folder,
+        )
+
         with (
-            maybe_enable_profiling(
-                job_config.profiling,
-                global_step=self.step,
-                base_folder=job_config.job.dump_folder,
-            ) as torch_profiler,
             maybe_enable_memory_snapshot(
                 job_config.profiling,
                 global_step=self.step,

torchtitan/tools/profiling.py

@@ -18,7 +18,6 @@
 MEMORY_SNAPSHOT_MAX_ENTRIES = 100000
 
 
-@contextlib.contextmanager
 def maybe_enable_profiling(
     profiling_config: ProfilingConfig,
     *,
@@ -68,20 +67,20 @@ def trace_handler(prof):
             gpu_device_profiled = torch.profiler.ProfilerActivity.CUDA
         elif torch.xpu.is_available():
             gpu_device_profiled = torch.profiler.ProfilerActivity.XPU
-        with torch.profiler.profile(
+        torch_profiler = torch.profiler.profile(
             activities=[
                 torch.profiler.ProfilerActivity.CPU,
                 gpu_device_profiled,
             ],
             schedule=torch.profiler.schedule(wait=wait, warmup=warmup, active=active),
             on_trace_ready=trace_handler,
             record_shapes=True,
-        ) as torch_profiler:
-            torch_profiler.step_num = global_step
-            yield torch_profiler
+        )
+        torch_profiler.step_num = global_step
+        torch_profiler.start()
+        return torch_profiler
     else:
-        torch_profiler = contextlib.nullcontext()
-        yield None
+        return None
 
 
 @contextlib.contextmanager

torchtitan/train.py

@@ -4,8 +4,10 @@
 # This source code is licensed under the BSD-style license found in the
 # LICENSE file in the root directory of this source tree.
 
+import ctypes
 import importlib
 import os
+import signal
 import time
 from datetime import timedelta
 from typing import Any, Generator, Iterable
@@ -33,8 +35,12 @@
     maybe_enable_profiling,
 )
 
+c_globals = ctypes.CDLL(None)  # POSIX
+
 
 class Trainer(torch.distributed.checkpoint.stateful.Stateful):
+    torch_profiler: torch.profiler.profile | None = None
+
     # core configs
     job_config: JobConfig
     parallel_dims: ParallelDims
@@ -613,13 +619,14 @@ def train(self):
             if not self.ft_manager.enabled
             else f"replica_{self.ft_manager.replica_id}"
         )
+        self.torch_profiler = maybe_enable_profiling(
+            job_config.profiling,
+            global_step=self.step,
+            base_folder=job_config.job.dump_folder,
+            leaf_folder=leaf_folder,
+        )
+
         with (
-            maybe_enable_profiling(
-                job_config.profiling,
-                global_step=self.step,
-                base_folder=job_config.job.dump_folder,
-                leaf_folder=leaf_folder,
-            ) as torch_profiler,
             maybe_enable_memory_snapshot(
                 job_config.profiling,
                 global_step=self.step,
@@ -643,6 +650,15 @@ def train(self):
                 ),
             ),
         ):
+            if self.torch_profiler:
+
+                @ctypes.CFUNCTYPE(None, ctypes.c_int)
+                def sigabrt_handler(signal):
+                    logger.info("SIGABRT received. Stopping profiler")
+                    self.torch_profiler.export_chrome_trace("trace.json")
+
+                c_globals.signal(signal.SIGABRT, sigabrt_handler)
+
             data_iterator = self.batch_generator(self.dataloader)
             while self.should_continue_training():
                 self.step += 1
@@ -666,8 +682,8 @@ def train(self):
                         self.validator.validate(self.model_parts, self.step)
 
                 # signal the profiler that the next profiling step has started
-                if torch_profiler:
-                    torch_profiler.step()
+                if self.torch_profiler:
+                    self.torch_profiler.step()
                 if memory_profiler:
                     memory_profiler.step()
 
@@ -730,10 +746,12 @@ def main(trainer_class: type[Trainer]) -> None:
         else:
             trainer.train()
     except Exception:
+        logger.info("Torchtitan training threw an exception")
         if trainer:
             trainer.close()
         raise
     else:
+        logger.info("Torchtitan training completed")
         trainer.close()
         torch.distributed.destroy_process_group()
         logger.info("Process group destroyed")