fix(redmule): unmap Conv from RedMulE engine; drop weight-layout pass

The Siracusa+RedMulE training CI on 1782a88 got past Python codegen but
failed at link time:

    ld.lld: error: undefined symbol:
        Conv2d_Im2Col_fp32_fp32_fp32_HWC_8_Redmule
    >>> referenced by TrainingNetwork.c:5386 in
        _node_1_tokenizer_..._Conv_cluster_fork

The original RedMulE PR (pulp-platform/Deeploy#67) shipped only the
matmul kernel TargetLibraries/PULPOpen/src/Matmul_fp32_Redmule.c.  The
ConvTemplate references a `Conv2d_Im2Col_..._8_Redmule` kernel that has
no corresponding source in the tree, and 67b754b already deleted the
testFloat2DConvolution / testFloat2dConvLarge fixtures that would have
exercised the Redmule Conv path.  So the Conv binding has always been
load-bearing only for non-test models like CCT_train, and on those it
breaks the link.

Two coupled changes route Conv through the existing PULPClusterEngine
(which has a working PULP_Conv2d_Im2Col_fp32_fp32_fp32_HWC):

- Drop 'Conv' from RedmuleMapping.  Without it Conv falls through to
  the second engine in RedmulePlatform's engine list (PULPCluster).
- Drop RedMuleAdjustWeightMemoryLayoutPass from the lowering passes.
  That pass transposed Conv weights from [F,H,W,Cin] to [H,W,Cin,F]
  for the RedMulE accelerator's expected layout; once Conv is on the
  PULPCluster engine, PULP expects [F,H,W,Cin] and the pre-applied
  transpose makes Tiling produce out-of-bounds tile rectangles
  (locally repro'd: AssertionError "Rectangle offset should be zero
  when the dimensions are the same. Received rectangle
  HyperRectangle(offset=(3, 0, 0, 0), dims=(3, 3, 3, 32))" in
  TilingCodegen.minimizeRectangle).

Both are clearly marked in-source as "restore when the RedMulE Conv
kernel lands."  Locally validated end-to-end:
- testMVPTraining.py    -> exit 0 (TrainingNetwork.c emits
  PULP_Conv2d_Im2Col_fp32_fp32_fp32_HWC for the tokenizer Conv).
- testMVPOptimizer.py   -> exit 0.

Matmul / Gemm continue to bind to RedMulE as before.

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>

Author: runwangdl

Deeploy/Targets/Redmule/Deployer.py

@@ -30,8 +30,7 @@
 from Deeploy.AbstractDataTypes import Pointer
 from Deeploy.DeeployTypes import DeploymentPlatform, TopologyOptimizer
 from Deeploy.Targets.PULPOpen.Deployer import PULPDeployer
-from Deeploy.Targets.Redmule.TopologyOptimizationPasses.Passes import RedMuleAdjustWeightMemoryLayoutPass, \
-    RedMuleGEMMTransposePass
+from Deeploy.Targets.Redmule.TopologyOptimizationPasses.Passes import RedMuleGEMMTransposePass
 
 
 class RedmuleDeployer(PULPDeployer):
@@ -50,6 +49,13 @@ def __init__(self,
                          default_channels_first, deeployStateDir, inputOffsets)
 
         self.loweringOptimizer.passes += [
-            RedMuleAdjustWeightMemoryLayoutPass("Redmule"),
+            # RedMuleAdjustWeightMemoryLayoutPass is intentionally not
+            # registered: it transposes Conv weights from [F,H,W,Cin] to
+            # [H,W,Cin,F] for the RedMulE accelerator, but Conv has been
+            # routed to the PULPCluster engine (see Engine.RedmuleMapping)
+            # because no `Conv2d_Im2Col_*_Redmule` kernel is defined yet.
+            # PULP's Conv expects [F,H,W,Cin] layout, so applying the
+            # transpose breaks tiling for those nodes.  Restore alongside
+            # the Conv mapping when a real RedMulE Conv kernel lands.
             RedMuleGEMMTransposePass("Redmule")
         ]

Deeploy/Targets/Redmule/Engine.py

@@ -39,7 +39,16 @@
 
 RedmuleMapping = {
     'MatMul': MatMulLayer([MatMulRedmuleMapper]),
-    'Conv': ConvLayer([Conv2DRedmuleMapper]),
+    # 'Conv' is intentionally not mapped here: the Redmule ConvTemplate
+    # references the kernel symbol Conv2d_Im2Col_fp32_fp32_fp32_HWC_8_Redmule,
+    # which is *declared* by the template but never *defined* in any source
+    # file under TargetLibraries/.  Letting Conv fall through to the next
+    # engine (PULPClusterEngine, which has a working
+    # PULP_Conv2d_Im2Col_fp32_fp32_fp32_HWC implementation) keeps the
+    # Siracusa+RedMulE link step from failing on undefined symbols.  When
+    # a real RedMulE-accelerated Conv kernel lands, restore the mapping:
+    #
+    #     'Conv': ConvLayer([Conv2DRedmuleMapper]),
     'Gemm': GEMMLayer([GEMMMRedmuleMapper]),
 }