[Refactor] MLP weight prefetch to consistency with MoE Model's prefetching in terms of code and usage (#6442)

### What this PR does / why we need it?
Refactor MLP weight prefetch to consistency with MoE Model's prefetching
in terms of code and usage.
Environments VLLM_ASCEND_ENABLE_PREFETCH_MLP,
VLLM_ASCEND_MLP_DOWN_PREFETCH_SIZE and
VLLM_ASCEND_MLP_GATE_UP_PREFETCH_SIZE is removed, usage as following:

--additional-config '{"weight_prefetch_config": { "enabled": true,
"prefetch_ratio": {"mlp": { "gate_up": 1.0, "down": 1.0} }}}'

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

### How was this patch tested?

- vLLM version: v0.14.1
- vLLM main:
dc917cceb8

---------

Signed-off-by: leo-pony <nengjunma@outlook.com>

docs/source/tutorials/Qwen3-Dense.md

@@ -171,9 +171,6 @@ export TASK_QUEUE_ENABLE=1
 # Enable the AIVector core to directly schedule ROCE communication
 export HCCL_OP_EXPANSION_MODE="AIV"
 
-# Enable MLP prefetch for better performance.
-export VLLM_ASCEND_ENABLE_PREFETCH_MLP=1
-
 # Enable FlashComm_v1 optimization when tensor parallel is enabled.
 export VLLM_ASCEND_ENABLE_FLASHCOMM1=1
 
@@ -187,7 +184,7 @@ vllm serve /model/Qwen3-32B-W8A8 \
   --max-model-len 5500 \
   --max-num-batched-tokens 40960 \
   --compilation-config '{"cudagraph_mode": "FULL_DECODE_ONLY"}' \
-  --additional-config '{"pa_shape_list":[48,64,72,80]}' \
+  --additional-config '{"pa_shape_list":[48,64,72,80], "weight_prefetch_config":{"enabled":true}}' \
   --port 8113 \
   --block-size 128 \
   --gpu-memory-utilization 0.9
@@ -348,9 +345,7 @@ Weight prefetching optimizes memory usage by preloading weights into the cache b
 
 In dense model scenarios, the MLP's gate_up_proj and down_proj linear layers often exhibit relatively high MTE utilization. To address this, we create a separate pipeline specifically for weight prefetching, which runs in parallel with the original vector computation pipeline, such as RMSNorm and SiLU, before the MLP. This approach allows the weights to be preloaded to L2 cache ahead of time, reducing MTE utilization during the MLP computations and indirectly improving Cube computation efficiency by minimizing resource contention and optimizing data flow.
 
-It is important to emphasize that, since we use vector computations to hide the weight prefetching pipeline, the setting of the prefetch buffer size is crucial. If the buffer size is too small, the optimization benefits will not be fully realized, while a larger buffer size may lead to resource contention, resulting in performance degradation. To accommodate different scenarios, we have exposed two environment variables `VLLM_ASCEND_MLP_GATE_UP_PREFETCH_SIZE` and `VLLM_ASCEND_MLP_DOWN_PREFETCH_SIZE` to allow for flexible buffer size configuration based on the specific workload.
-
-This optimization requires setting the environment variable `VLLM_ASCEND_ENABLE_PREFETCH_MLP = 1` to be enabled.
+Previously, the environment variables VLLM_ASCEND_ENABLE_PREFETCH_MLP used to enable MLP weight prefetch and VLLM_ASCEND_MLP_GATE_UP_PREFETCH_SIZE and VLLM_ASCEND_MLP_DOWN_PREFETCH_SIZE used to set the weight prefetch size for MLP gate_up_proj and down_proj were deprecated. Please use the following configuration instead: "weight_prefetch_config": { "enabled": true, "prefetch_ratio": { "mlp": { "gate_up": 1.0, "down": 1.0}}}. See User Guide->Feature Guide->Weight Prefetch Guide for details.
 
 ### 6. Zerolike Elimination
 

docs/source/user_guide/configuration/additional_config.md

@@ -60,7 +60,7 @@ The details of each configuration option are as follows:
 | Name             | Type | Default                                                     | Description                        |
 |------------------|------|-------------------------------------------------------------|------------------------------------|
 | `enabled`        | bool | `False`                                                     | Whether to enable weight prefetch. |
-| `prefetch_ratio` | dict | `{"attn": {"qkv": 1.0, "o": 1.0}, "moe": {"gate_up": 0.8}}` | Prefetch ratio of each weight.     |
+| `prefetch_ratio` | dict | `{"attn": {"qkv": 1.0, "o": 1.0}, "moe": {"gate_up": 0.8}, "mlp": { "gate_up": 1.0,  "down": 1.0}}` | Prefetch ratio of each weight.     |
 
 **finegrained_tp_config**
 
@@ -115,6 +115,10 @@ An example of additional configuration is as follows:
             },
             "moe": {
                 "gate_up": 0.8
+            },
+            "mlp": {
+                "gate_up": 1.0,
+                "down": 1.0
             }
         },
     },

docs/source/user_guide/feature_guide/index.md

@@ -23,4 +23,5 @@ layer_sharding
 speculative_decoding
 context_parallel
 npugraph_ex
+weight_prefetch
 :::

docs/source/user_guide/feature_guide/weight_prefetch.md

@@ -0,0 +1,73 @@
+# Weight Prefetch Guide
+
+Weight prefetching optimizes memory usage by preloading weights into the cache before they are needed, minimizing delays caused by memory access during model execution. Linear layers sometimes exhibit relatively high MTE utilization. To address this, we create a separate pipeline specifically for weight prefetching, which runs in parallel with the original vector computation pipeline, such as quantize, MoE gating top_k, RMSNorm and SwiGlu. This approach allows the weights to be preloaded to L2 cache ahead of time, reducing MTE utilization during the linear layer computations and indirectly improving Cube computation efficiency by minimizing resource contention and optimizing data flow.
+
+Since we use vector computations to hide the weight prefetching pipeline, it has effect on computation, if you prioritize low latency over high throughput, then it it best not to enable prefetching.
+
+## Quick Start
+
+With `--additional-config '{"weight_prefetch_config": {"enabled": true}}'` to open weight prefetch.
+
+## Fine-tune Prefetch Ratio
+
+Since weight prefetch use vector computations to hide the weight prefetching pipeline, the setting of the prefetch size is crucial. If the size is too small, the optimization benefits will not be fully realized, while a larger size may lead to resource contention, resulting in performance degradation. To accommodate different scenarios, we have add  `prefetch_ratio` to allow for flexible size configuration based on the specific workload, detail as following:
+
+With `prefetch_ratio` in `"weight_prefetch_config"` to custom the weight prefetch ratio for specify linear layers.
+
+The “attn” and “moe” configuration options are used for MoE model, detail as following:
+
+`"attn": { "qkv": 1.0,  "o": 1.0},  "moe": {"gate_up": 0.8}`
+
+The “mlp” configuration option is used to optimize the performance of the Dense model, detail as following:
+
+ `"mlp": {"gate_up": 1.0, "down": 1.0}`
+
+Above value are the default config, the default value has a good performance for Qwen3-235B-A22B-W8A8 when `--max-num-seqs`is 144, for Qwen3-32B-W8A8 when `--max-num-seqs`is 72.
+
+However, this may not be the optimal configuration for your scenario. For higher concurrency, you can try increasing the prefetch size. For lower concurrency, prefetching may not offer any advantages, so you can decrease the size or disable prefetching. Determine if the prefetch size is appropriate by collecting profiling data. Specifically, check if the time required for the prefetch operation (e.g., MLP Down Proj weight prefetching) overlaps with the time required for parallel vector computation operators (e.g., SwiGlu computation), and whether the prefetch operation is no later than the completion time of the vector computation operator. In the profiling timeline, a prefetch operation appears as a CMO operation on a single stream; this CMO operation is the prefetch operation.
+
+Notices:
+
+1) Weight prefetch of MLP `down` project prefetch dependence sequence parallel, if you want open for mlp `down` please also enable sequence parallel.
+2) Due to the current size of the L2 cache, the maximum prefetch cannot exceed 18MB. If `prefetch_ration * lineaer_layer_weight_size >= 18 * 1024 * 1024` bytes, the backend will only prefetch 18MB.
+
+## Example
+
+1) For MoE model:
+
+```shell
+    --additional-config \
+    '{
+        "weight_prefetch_config": {
+            "enabled": true,
+            "prefetch_ratio": {
+                "attn": {
+                    "qkv": 1.0,
+                    "o": 1.0
+                },
+                "moe": {
+                    "gate_up": 0.8
+                }
+            }
+        }
+    }'
+```
+
+2) For dense model:
+
+Following is the default configuration that can get a good performance for `--max-num-seqs`is 72 for Qwen3-32B-W8A8
+
+```shell
+    --additional-config \
+    '{
+        "weight_prefetch_config": {
+            "enabled": true,
+            "prefetch_ratio": {
+                "mlp": {
+                    "gate_up": 1.0,
+                    "down": 1.0
+                }
+            }
+        }
+    }'
+```