[Doc] Add layer_sharding additional config for DeepSeek-V3.2-W8A8 (#5921)

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

#### Documentation Improvements

New Configuration: Added the layer_sharding parameter to the
DeepSeek-V3.2-W8A8 deployment tutorial. This guides users to include
`["q_b_proj", "o_proj"]` in their prefill node setup for better resource
utilization.

#### CI and Testing Updates

Test Config Update: Updated the multi-node E2E test configuration file:
tests/e2e/nightly/multi_node/config/DeepSeek-V3_2-W8A8-A3-dual-nodes.yaml.

including disable `FLASHCOMM` and enable `FULL_DECODE_ONLY` and update
performance baseline.

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

Yes. The documentation now recommends a more optimized startup command
for DeepSeek-V3.2-W8A8. Users following the updated tutorial will see
improved performance in multi-node PD disaggregation environments.

### How was this patch tested?
CI Validation: The updated E2E test configuration has been verified
through the nightly CI pipeline.

Environment: * vLLM version: v0.13.0

Base Commit:
[11b6af5](11b6af5280)

Hardware: Ascend A3/A2 multi-node cluster.

---------

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

docs/source/tutorials/DeepSeek-V3.2.md

@@ -313,6 +313,7 @@ Before you start, please
             --quantization ascend \
             --enforce-eager \
             --no-enable-prefix-caching \
+            --additional-config '{"layer_sharding": ["q_b_proj", "o_proj"]}' \
             --kv-transfer-config \
             '{"kv_connector": "MooncakeConnectorV1",
             "kv_role": "kv_producer",
@@ -388,6 +389,7 @@ Before you start, please
             --quantization ascend \
             --enforce-eager \
             --no-enable-prefix-caching \
+            --additional-config '{"layer_sharding": ["q_b_proj", "o_proj"]}' \
             --kv-transfer-config \
             '{"kv_connector": "MooncakeConnectorV1",
             "kv_role": "kv_producer",

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

@@ -12,7 +12,7 @@ env_common:
   OMP_NUM_THREADS: 1
   VLLM_ASCEND_ENABLE_MLAPO: 1
   PYTORCH_NPU_ALLOC_CONF: "expandable_segments:True"
-  VLLM_ASCEND_ENABLE_FLASHCOMM1: 1
+  VLLM_ASCEND_ENABLE_FLASHCOMM1: 0
   ASCEND_A3_EBA_ENABLE: 1
 
 
@@ -37,9 +37,9 @@ deployment:
       --gpu-memory-utilization 0.85
       --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"}' 
       --tokenizer-mode deepseek_v32
       --reasoning-parser deepseek_v3
-      --api-server-count 4
 
   -
     server_cmd: >
@@ -61,6 +61,7 @@ deployment:
       --gpu-memory-utilization 0.85
       --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"}' 
       --tokenizer-mode deepseek_v32
       --reasoning-parser deepseek_v3
 benchmarks:
@@ -73,7 +74,7 @@ benchmarks:
     max_out_len: 3000
     batch_size: 512
     request_rate: 11.2
-    baseline: 594.915
+    baseline: 905.6805
     threshold: 0.97
   acc:
     case_type: accuracy