Fix group fusion stride layout (pytorch#2442)

Summary:
Pull Request resolved: pytorch#2442

context:
https://fb.workplace.com/groups/1075192433118967/permalink/1401282167176657/

moving the changes to the group gemm op has compilation errors, see details in D55606636

Differential Revision: D61888433

userbenchmark/dynamo/dynamobench/_dynamo/utils.py

@@ -3242,3 +3242,9 @@ def record(cls):
         finally:
             if config.record_compile_time_instruction_count:
                 cls.end()
+
+
+def realize_inputs(inputs: List[torch.fx.Node]):
+    for inp in inputs:
+        if isinstance(inp, torch.fx.node.Node):
+            inp.meta["inductor_realize_to_strides"] = True