[CI] Enable FLASHCOMM1 with layer_sharding and FULL_DECODE_ONLY in ds32 testing (#6115)

### What this PR does / why we need it?

This PR enables FLASHCOMM1 communication optimization with layer
sharding for DeepSeek-V3.2 W8A8 model testing to
  validate PR #5702. The changes include:

  1. Enable FLASHCOMM1: Set VLLM_ASCEND_ENABLE_FLASHCOMM1=1
  improves performance for distributed inference
2. Add layer sharding: Configure layer_sharding: ["q_b_proj", "o_proj"]
4. Update baselines: Adjust performance baselines to reflect the
improvements from FLASHCOMM1 and layer sharding

### Does this PR introduce _any_ user-facing change?

No. This is a CI/test-only change that enables new communication
optimization features for testing purposes.

### How was this patch tested?

- vLLM version: v0.13.0
- vLLM main:
d68209402d

Signed-off-by: guozr <guozr1997@hotmail.com>
Co-authored-by: guozr <guozr1997@hotmail.com>

tests/e2e/multicard/2-cards/test_offline_inference_distributed.py

@@ -242,7 +242,7 @@ def test_qwen3_dense_prefetch_mlp_weight_tp2(model):
 
 
 @patch.dict(os.environ, {"HCCL_OP_EXPANSION_MODE": "AIV"})
-@patch.dict(os.environ, {"VLLM_ASCEND_ENABLE_FLASHCOMM1": "0"})
+@patch.dict(os.environ, {"VLLM_ASCEND_ENABLE_FLASHCOMM1": "1"})
 @patch.dict(os.environ, {"ASCEND_AGGREGATE_ENABLE": "1"})
 @patch.dict(os.environ, {"HCCL_BUFFSIZE": "1024"})
 def test_deepseek3_2_w8a8_pruning_mtp_tp2_ep():
@@ -262,6 +262,9 @@ def test_deepseek3_2_w8a8_pruning_mtp_tp2_ep():
                         "num_speculative_tokens": 2,
                         "method": "deepseek_mtp"
                     },
+                    additional_config={
+                        "layer_sharding":["q_b_proj", "o_proj"]
+                    },
                     reasoning_parser="deepseek_v3",
                     tokenizer_mode="deepseek_v32") as vllm_model:
         vllm_model.generate_greedy(example_prompts, max_tokens)

tests/e2e/nightly/multi_node/config/DeepSeek-V3_2-W8A8-A3-dual-nodes.yaml

@@ -11,7 +11,7 @@ env_common:
   OMP_PROC_BIND: false
   OMP_NUM_THREADS: 1
   PYTORCH_NPU_ALLOC_CONF: "expandable_segments:True"
-  VLLM_ASCEND_ENABLE_FLASHCOMM1: 0
+  VLLM_ASCEND_ENABLE_FLASHCOMM1: 1
   ASCEND_A3_EBA_ENABLE: 1
 
 
@@ -37,6 +37,7 @@ deployment:
       --trust-remote-code
       --speculative-config '{"num_speculative_tokens": 2, "method":"deepseek_mtp"}'
       --compilation-config '{"cudagraph_capture_sizes": [3, 6, 9, 12, 15, 18, 21, 24, 27, 30, 33, 36, 39, 42, 45, 48], "cudagraph_mode": "FULL_DECODE_ONLY"}' 
+      --additional-config '{"layer_sharding": ["q_b_proj", "o_proj"]}'
       --tokenizer-mode deepseek_v32
       --reasoning-parser deepseek_v3
 
@@ -61,6 +62,7 @@ deployment:
       --trust-remote-code
       --speculative-config '{"num_speculative_tokens": 2, "method":"deepseek_mtp"}'
       --compilation-config '{"cudagraph_capture_sizes": [3, 6, 9, 12, 15, 18, 21, 24, 27, 30, 33, 36, 39, 42, 45, 48], "cudagraph_mode": "FULL_DECODE_ONLY"}' 
+      --additional-config '{"layer_sharding": ["q_b_proj", "o_proj"]}'
       --tokenizer-mode deepseek_v32
       --reasoning-parser deepseek_v3
 benchmarks:
@@ -73,8 +75,9 @@ benchmarks:
     max_out_len: 3000
     batch_size: 512
     request_rate: 11.2
-    baseline: 905.6805
+    baseline: 1253.8466 
     threshold: 0.97
+    
   acc:
     case_type: accuracy
     dataset_path: vllm-ascend/gsm8k-lite

tests/e2e/nightly/single_node/models/test_deepseek_v3_2_w8a8.py

@@ -54,7 +54,7 @@ aisbench_cases = [{
     "max_out_len": 1500,
     "batch_size": 4,
     "request_rate": 11.2,
-    "baseline": 120,
+    "baseline": 134,
     "threshold": 0.97
 }]
 
@@ -72,7 +72,7 @@ async def test_models(model: str, tp_size: int, dp_size: int) -> None:
         "HCCL_BUFFSIZE": "1024",
         "VLLM_ASCEND_ENABLE_MLAPO": "1",
         "PYTORCH_NPU_ALLOC_CONF": "expandable_segments:True",
-        "VLLM_ASCEND_ENABLE_FLASHCOMM1": "0",
+        "VLLM_ASCEND_ENABLE_FLASHCOMM1": "1",
     }
 
     server_args = [
@@ -85,6 +85,8 @@ async def test_models(model: str, tp_size: int, dp_size: int) -> None:
         '{"cudagraph_capture_sizes":[3, 6, 9, 12], "cudagraph_mode":"FULL_DECODE_ONLY"}',
         "--speculative-config",
         '{"num_speculative_tokens": 2, "method":"deepseek_mtp"}',
+        "--additional-config",
+        '{"layer_sharding": ["q_b_proj", "o_proj"]}',
         "--reasoning-parser", "deepseek_v3", "--tokenizer_mode", "deepseek_v32"
     ]
     request_keyword_args: dict[str, Any] = {