[BE] replace the extra DeviceMesh _flatten with mesh access

ghstack-source-id: a0689ec03803419d67a4a79ec325dfed15113cdf
Pull Request resolved: #666

torchtitan/parallelisms/parallelize_llama.py

@@ -34,6 +34,7 @@
 from torchtitan.config_manager import JobConfig, TORCH_DTYPE_MAP
 from torchtitan.logging import logger
 from torchtitan.parallelisms.parallel_dims import ParallelDims
+from torchtitan.parallelisms.utils import check_if_feature_in_pytorch
 
 
 def parallelize_llama(
@@ -79,22 +80,31 @@ def parallelize_llama(
     if (
         parallel_dims.dp_shard_enabled
     ):  # apply FSDP or HSDP, potentially with Context Parallel
-
-        # TODO: instead of flattening the mesh twice, we could've done in a batter way:
-        # dp_mesh = world_mesh["dp_cp"] if parallel_dims.cp_enabled else world_mesh["dp"]
-        # However, this leads to an error in `DeviceMesh.__get_item__` which I believe is
-        # a bug in DeviceMesh. We should fix it and then use the above line.
-        dp_mesh_dim_names = (
-            ("dp_replicate", "dp_shard")
-            if parallel_dims.dp_replicate_enabled
-            else ("dp",)
-        )
-        # note that mesh can only be flattened from the finest-grained mesh dimensions
-        dp_mesh = (
-            world_mesh[(*dp_mesh_dim_names, "cp")]._flatten("dp_cp")
-            if parallel_dims.cp_enabled
-            else world_mesh[dp_mesh_dim_names]
-        )
+        try:
+            dp_mesh = (
+                world_mesh["dp_cp"] if parallel_dims.cp_enabled else world_mesh["dp"]
+            )
+        except IndexError:
+            # note: this is a workaround of the above logic for old pytorch version
+            # where https://github.com/pytorch/pytorch/pull/138945 is not included
+            # throw a warning to encourage users to upgrade to a newer pytorch version
+            check_if_feature_in_pytorch(
+                "DeviceMesh flattening over 3D+ meshes",
+                "https://github.com/pytorch/pytorch/pull/138945",
+                "2.6.0.dev20241030",
+            )
+            # TODO: remove this workaround once PyTorch 2.6 is released
+            dp_mesh_dim_names = (
+                ("dp_replicate", "dp_shard")
+                if parallel_dims.dp_replicate_enabled
+                else ("dp",)
+            )
+            # note that mesh can only be flattened from the finest-grained mesh dimensions
+            dp_mesh = (
+                world_mesh[(*dp_mesh_dim_names, "cp")]._flatten("dp_cp")
+                if parallel_dims.cp_enabled
+                else world_mesh[dp_mesh_dim_names]
+            )
 
         apply_fsdp(
             model,

torchtitan/parallelisms/utils.py

@@ -0,0 +1,28 @@
+# Copyright (c) Meta Platforms, Inc. and affiliates.
+# All rights reserved.
+#
+# This source code is licensed under the BSD-style license found in the
+# LICENSE file in the root directory of this source tree.
+from typing import Optional
+
+import torch
+from torchtitan.logging import logger
+
+
+def check_if_feature_in_pytorch(
+    feature_name: str,
+    pull_request: str,
+    min_nightly_version: Optional[str] = None,
+) -> None:
+    if "git" in torch.__version__:  # pytorch is built from source
+        # notify users to check if the pull request is included in their pytorch
+        logger.warning(
+            "detected that the pytorch is built from source. Please make sure the PR "
+            f"({pull_request_link}) is included in pytorch for correct {feature_name}."
+        )
+    elif min_nightly_version is not None and torch.__version__ < min_nightly_version:
+        logger.warning(
+            f"detected that the pytorch version {torch.__version__} is older than "
+            f"{min_nightly_version}. Please upgrade a newer version to include the "
+            f"change in ({pull_request_link}) for correct {feature_name}."
+        )