diff --git a/computility-run.yaml b/computility-run.yaml
index def56434..015fca7f 100644
--- a/computility-run.yaml
+++ b/computility-run.yaml
@@ -48,8 +48,6 @@ env:
     - name: BI100_MOE_COREX_TOPK_SOFTMAX
       value: '1'
     - name: PYTORCH_CUDA_ALLOC_CONF
-      value: max_split_size_mb:512
+      value: expandable_segments:True
     - name: LD_PRELOAD
-      value: /workspace/qwen3_6_scripts/libcccl_allocator.so
-    - name: CCCL_ALLOC_DISABLE
-      value: '0'
+      value: /workspace/qwen3_6_scripts/cccl_preload/libcccl_allocator.so
diff --git a/qwen3_6_scripts/patch_ops.sh b/qwen3_6_scripts/patch_ops.sh
index 2c3f6585..28a794c0 100755
--- a/qwen3_6_scripts/patch_ops.sh
+++ b/qwen3_6_scripts/patch_ops.sh
@@ -247,15 +247,8 @@ if source != installed:
 PY
 
 build_stage "compiling CCCL CachingDeviceAllocator LD_PRELOAD module"
-if bash ./cccl_preload/build_cccl_preload.sh /workspace/qwen3_6_scripts; then
-    if [[ -s /workspace/qwen3_6_scripts/libcccl_allocator.so ]]; then
-        echo "[OK] libcccl_allocator.so built successfully"
-    else
-        echo "[WARN] libcccl_allocator.so is empty or missing after build"
-    fi
-else
-    echo "[WARN] CCCL preload allocator build failed — LD_PRELOAD will be ignored at runtime"
-fi
+bash ./cccl_preload/build_cccl_preload.sh /workspace/qwen3_6_scripts/cccl_preload || \
+    echo "[WARN] CCCL preload allocator build failed — will use default allocator"
 
 build_stage "compiling CoreX CUDA extensions (moe_index_combine + gdn_chunk_recurrent)"
 if [[ -x /usr/local/corex-3.2.3/bin/clang++ ]]; then