Phase-3: remove len<=1 guard in promotedReferenceReset

Allows the absolute offset to fire for single-tile transfers too.
Does not fix CCT_2 activation promotion (32KB activations have
1 L2->L1 tile with offset=0, the spatial slicing was handled by
the now-absent L3 closure).

Author: runwangdl

Deeploy/TilingExtension/CodeTransformationPasses/TilingCodeGeneration.py

@@ -160,8 +160,6 @@ def _generatePromotedReferenceReset(self, ctxt: NetworkContext, tensorName: str,
         body top (which resets every call).  Covers both the read side
         (L2→L1 for promoted constants) and the write side (L1→L2 for
         promoted activations)."""
-        if len(transfers) <= 1:
-            return None
         from Deeploy.DeeployTypes import TransientBuffer
         try:
             unraveled = ctxt.unravelReference(externalBuffer)