From 5e4b00d62020db20f49371e86793aa6acd709de2 Mon Sep 17 00:00:00 2001
From: Masaki Kozuki <mkozuki@nvidia.com>
Date: Wed, 6 Nov 2024 15:31:44 +0900
Subject: [PATCH] handle `PrimIDs.RETURN` earlier

Signed-off-by: Masaki Kozuki <mkozuki@nvidia.com>
---
 thunder/transforms/tensor_subclasses.py | 62 ++++++++++++-------------
 1 file changed, 31 insertions(+), 31 deletions(-)

diff --git a/thunder/transforms/tensor_subclasses.py b/thunder/transforms/tensor_subclasses.py
index 677e6f67ec..979bdf31e5 100644
--- a/thunder/transforms/tensor_subclasses.py
+++ b/thunder/transforms/tensor_subclasses.py
@@ -450,6 +450,37 @@ def f_with_wrap_and_unwrap(*desugared_args) -> tuple[OutputWrapperForFxTracing,
 
     def __call__(self, bsym: BoundSymbol) -> list[BoundSymbol]:
         updated_bsym: BoundSymbol = bsym.from_bsym_swap_proxies(self.swap_map)
+        if updated_bsym.sym.id == prims.PrimIDs.RETURN:
+            unflatten_fake_tensor_result = tree_unflatten(
+                self.fx_computation_trace_result,
+                self.spec_of_fx_computation_trace_result,
+            )
+            outputs: dict[str, Any] = updated_bsym.args[0]  # {"output": ..., "flat_args": ...}
+            utils.check_type(outputs, dict)
+            utils.check(
+                isinstance(outputs, dict) and len(outputs) == 2 and ("output", "flat_args") == tuple(outputs.keys()),
+                lambda: fr"{outputs=} does not conform to the format of \{'output': ..., 'flat_args': [...]\}",
+            )
+            seq_outs = utils.sequencify(outputs["output"])
+            seq_fake_ret = utils.sequencify(unflatten_fake_tensor_result["output"])
+            utils.check(
+                len(seq_outs) == len(seq_fake_ret),
+                lambda: f"{outputs['output']=}, {unflatten_fake_tensor_result['output']=}",
+            )
+
+            bsyms: list[BoundSymbol] = []
+            for proxy_output, fx_output in zip(seq_outs, seq_fake_ret):
+                if not isinstance(proxy_output, SubclassTensorProxy):
+                    continue
+                tensor_attrs, metadata = proxy_output.__tensor_flatten__()
+                tensors = [getattr(proxy_output, name) for name in tensor_attrs]
+                bsyms.append(
+                    prims.unflatten_tensor_subclass.bind(
+                        type(fx_output), dict(zip(tensor_attrs, tensors)), metadata, output=proxy_output
+                    )
+                )
+            return [*bsyms, updated_bsym]
+
         if not any(isinstance(a, SubclassTensorProxy) for a in updated_bsym.flat_proxy_args):
             if bsym.sym.id == prims.PrimIDs.TENSOR_SUBCLASS_CTOR:
                 subclass_proxy = updated_bsym.flat_proxy_outs[0]
@@ -483,37 +514,6 @@ def __call__(self, bsym: BoundSymbol) -> list[BoundSymbol]:
             exception_type=NotImplementedError,
         )
 
-        if updated_bsym.sym.id == prims.PrimIDs.RETURN:
-            unflatten_fake_tensor_result = tree_unflatten(
-                self.fx_computation_trace_result,
-                self.spec_of_fx_computation_trace_result,
-            )
-            outputs: dict[str, Any] = updated_bsym.args[0]  # {"output": ..., "flat_args": ...}
-            utils.check_type(outputs, dict)
-            utils.check(
-                isinstance(outputs, dict) and len(outputs) == 2 and ("output", "flat_args") == tuple(outputs.keys()),
-                lambda: fr"{outputs=} does not conform to the format of \{'output': ..., 'flat_args': [...]\}",
-            )
-            seq_outs = utils.sequencify(outputs["output"])
-            seq_fake_ret = utils.sequencify(unflatten_fake_tensor_result["output"])
-            utils.check(
-                len(seq_outs) == len(seq_fake_ret),
-                lambda: f"{outputs['output']=}, {unflatten_fake_tensor_result['output']=}",
-            )
-
-            bsyms: list[BoundSymbol] = []
-            for proxy_output, fx_output in zip(seq_outs, seq_fake_ret):
-                if not isinstance(proxy_output, SubclassTensorProxy):
-                    continue
-                tensor_attrs, metadata = proxy_output.__tensor_flatten__()
-                tensors = [getattr(proxy_output, name) for name in tensor_attrs]
-                bsyms.append(
-                    prims.unflatten_tensor_subclass.bind(
-                        type(fx_output), dict(zip(tensor_attrs, tensors)), metadata, output=proxy_output
-                    )
-                )
-            return [*bsyms, updated_bsym]
-
         trace = trace_from_bsym_or_bsyms(updated_bsym)
         fx, sequencified_cosmeticized_out, orig_output, _ = self.convert_trace_to_fx_graph_and_get_fake_result(trace)
         utils.check(