feat: Cu10 TensorOp batched HGEMM via Iluvatar CUTLASS framework

From corex-samples batched_gemm.cu, changed:
  float → half_t, OpClassSimt → OpClassTensorOp, Sm61 → Cu10
Uses __ivcorex_matrix_mad_f32x4_f16x4 via mma_cu10.h
Default config: TB<128,128,32> Warp<32,32,32> Inst<16,16,16>

Standalone test: correctness + perf for MoE decode (8 × 1x4096@4096x11008)

ex_engine/xllm_kernels/build_test_cutlass_batched.sh

@@ -0,0 +1,25 @@
+#!/bin/bash
+# build_test_cutlass_batched.sh — Compile and test Cu10 TensorOp batched GEMM
+set -eo pipefail
+
+SAMPLES="/usr/local/corex-samples-3.2.3_x86_64/samples/cutlass"
+SRC="ex_engine/xllm_kernels/cuda/moe_cutlass_batched.cu"
+
+echo "=== Compile Cu10 TensorOp batched HGEMM ==="
+/usr/local/corex/bin/clang++ \
+    --cuda-gpu-arch=ivcore10 --cuda-path=/usr/local/corex \
+    -I"${SAMPLES}/include" \
+    -I/usr/local/corex/include \
+    -L/usr/local/corex/lib64 -lcudart -lcutlass \
+    -DBUILD_STANDALONE_TEST \
+    -O2 -std=c++17 \
+    "$SRC" -o /tmp/test_cutlass_batched 2>&1
+
+if [ -f /tmp/test_cutlass_batched ]; then
+    echo "Compile: SUCCESS"
+    echo ""
+    echo "=== Run ==="
+    /tmp/test_cutlass_batched
+else
+    echo "Compile: FAILED"
+fi