[main] add pd transfer for ascend scheduler (#2753)

### What this PR does / why we need it?
For offline scenarios, adjust the scheduling process to prioritize the
prefill phase of all requests, then process the decode phase of all
requests.

### How was this patch tested?

```
max_num_seqs=24,
additional_config={
    "ascend_scheduler_config":{
        "enabled": True,
        "enable_pd_transfer": True,
        "decode_max_num_seqs": 24,
        "enable_chunked_prefill": False
    }
},
```
| input | output | num prompts | max_num_seqs | dp | tp | scheduler |
tps |
| ------ | ------ | ---------- | ---------------- | ---- | ---- |
---------------- | --------------- |
| dapo-math-17K | 2K | 384 | 24 | 2 | 1 | v1 | 234.06 |
| dapo-math-17K | 2K | 384 | 24 | 2 | 1 | pd transfer | 239.59(+2.4%) |
| dapo-math-17K| 2K | 384 | 24 | 4 | 1 | v1 | 222.85 |
| dapo-math-17K| 2K | 384 | 24 | 4 | 1 | pd transfer | 225.81(+1.3%) |


- vLLM version: v0.10.1.1
- vLLM main:
6fb2788163

---------

Signed-off-by: CaranLic <740821011@qq.com>

vllm_ascend/worker/model_runner_v1.py

@@ -66,7 +66,6 @@ from vllm.v1.kv_cache_interface import (FullAttentionSpec, KVCacheConfig,
 from vllm.v1.outputs import (EMPTY_MODEL_RUNNER_OUTPUT, DraftTokenIds,
                              LogprobsTensors, ModelRunnerOutput)
 from vllm.v1.pool.metadata import PoolingMetadata
-from vllm.v1.sample.logits_processor import build_logitsprocs
 from vllm.v1.sample.metadata import SamplingMetadata
 from vllm.v1.spec_decode.metadata import SpecDecodeMetadata
 from vllm.v1.spec_decode.ngram_proposer import NgramProposer
@@ -86,6 +85,7 @@ from vllm_ascend.attention.utils import AscendCommonAttentionMetadata
 from vllm_ascend.compilation.acl_graph import ACLGraphWrapper
 from vllm_ascend.multistream.ms_split import compute_split_seq_index
 from vllm_ascend.platform import NPUPlatform
+from vllm_ascend.sample.logits_processor import build_logitsprocs
 from vllm_ascend.sample.rejection_sampler import AscendRejectionSampler
 from vllm_ascend.spec_decode import get_spec_decode_method
 from vllm_ascend.spec_decode.eagle_proposer import EagleProposer
@@ -173,7 +173,10 @@ class NPUModelRunner(LoRAModelRunnerMixin):
         self.max_num_blocks_per_req = cdiv(self.model_config.max_model_len,
                                            self.block_size)
         self.max_num_tokens = self.scheduler_config.max_num_batched_tokens
-        self.max_num_reqs = self.scheduler_config.max_num_seqs
+        decode_max_num_seqs = getattr(self.scheduler_config,
+                                      'decode_max_num_seqs', 0)
+        self.max_num_reqs = max(self.scheduler_config.max_num_seqs,
+                                decode_max_num_seqs)
         self.dp_size = vllm_config.parallel_config.data_parallel_size
         self.dp_rank = vllm_config.parallel_config.data_parallel_rank
         self.device = device