fix(EX): corex ivcore10 build flags + deploy pipeline + topk kernel cleanup

Real machine log (2d5232c dockerrizhi.txt) shows two AST call chain breaks:

1. EVERY layer EVERY token:
   _custom_ops.py:58 'ixformer.functions has no attribute vllm_moe_topk_softmax'
   -> FusedMoE falls to PyTorch loop (2304 calls/token)

2. EVERY GDN layer (4 layers):
   'NaN in prefill GatedDeltaNet layer N (frac=0.9998-1.0000)'
   -> _torch_chunk_gated_delta_rule produces all-NaN

Fixes:
- build.sh: --cuda-gpu-arch=ivcore10, -D__ILUVATAR__ flags from real log
- Dockerfile: add ex_engine build before patch_ops
- patch_ops.sh: deploy .so + python into vllm model dir
- ex_loader.py: search co-located .so paths
- patch_model.py: remove premature auto-apply
- factor_moe_topk_softmax.cu: remove dead parallel branch

ex_engine/python/patch_model.py

@@ -191,11 +191,7 @@ def _patch_gdn_prefill(engine):
 
 
 # ---------------------------------------------------------------------------
-# Auto-apply on import if build dir exists
+# Call apply_patches() explicitly AFTER vllm model modules are loaded.
+# Integration point: qwen3_5.py calls this at the end of model __init__,
+# or patch_ops.sh adds it to the startup sequence.
 # ---------------------------------------------------------------------------
-_AUTO_BUILD_DIR = os.environ.get("EX_ENGINE_BUILD_DIR", "/workspace/ex_engine/build")
-if os.path.isdir(_AUTO_BUILD_DIR):
-    try:
-        apply_patches(_AUTO_BUILD_DIR)
-    except Exception as e:
-        logger.warning("EX Engine auto-apply failed: %s", e)