Test eagle OOM

Signed-off-by: Icey <[email protected]>

Author: wxsIcey

.github/workflows/_e2e_test.yaml

@@ -83,6 +83,9 @@ jobs:
           VLLM_WORKER_MULTIPROC_METHOD: spawn
           VLLM_USE_MODELSCOPE: True
           PYTORCH_NPU_ALLOC_CONF: max_split_size_mb:256
+          PYTORCH_NPU_ALLOC_CONF: expandable_segments:True
+          NPU_MEMORY_FRACTION: 0.96
+          OMP_NUM_THREADS: 1
         if: ${{ inputs.type == 'full' }}
         run: |
           # We found that if running aclgraph tests in batch, it will cause AclmdlRICaptureBegin error. So we run
@@ -103,8 +106,7 @@ jobs:
           # ------------------------------------ v1 spec decode test ------------------------------------ #
           pytest -sv tests/e2e/singlecard/spec_decode_v1/test_v1_mtp_correctness.py
           pytest -sv tests/e2e/singlecard/spec_decode_v1/test_v1_mtp_torchair_correctness.py
-          # Fix me: OOM error
-          #pytest -sv tests/e2e/singlecard/spec_decode_v1/test_v1_spec_decode.py
+          pytest -sv tests/e2e/singlecard/spec_decode_v1/test_v1_spec_decode.py
 
           pytest -sv tests/e2e/singlecard/ops/