[BugFix][0.18.0][310p] fix post-sampling not working in graph mode on 310p (#8077)

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

Enabling temperature in post-processing on 310P devices can cause the
service to stall and eventually hang. We first traced the issue to a
timeout where the temperature-related `div` operator was waiting for
results from a sub-stream. After investigating the preceding operators,
we finally identified the root cause as the `q.exponential_()` operator,
which is not well supported on 310P and triggers an internal issue in
the `add` kernel.

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

### How was this patch tested?
This patch was thoroughly tested locally（accuracy-dataset test and
stress test）. It is not easy to design a proper unit test for this case,
and I appreciate your understanding.

Signed-off-by: Tflowers-0129 <2906339855@qq.com>

vllm_ascend/_310p/model_runner_310p.py

@@ -35,7 +35,9 @@ from vllm.v1.kv_cache_interface import (
     MambaSpec,
     UniformTypeKVCacheSpecs,
 )
+from vllm.v1.sample.rejection_sampler import RejectionSampler
 
+from vllm_ascend._310p.sample.sampler import AscendSampler310
 from vllm_ascend.attention.attention_v1 import AscendAttentionState
 from vllm_ascend.utils import ACL_FORMAT_FRACTAL_NZ
 from vllm_ascend.worker.model_runner_v1 import NPUModelRunner
@@ -52,6 +54,9 @@ class NPUModelRunner310(NPUModelRunner):
     def __init__(self, *args, **kwargs):
         super().__init__(*args, **kwargs)
         self._acl_format = ACL_FORMAT_FRACTAL_NZ
+        self.sampler = AscendSampler310()
+        if getattr(self, "rejection_sampler", None) is not None:
+            self.rejection_sampler = RejectionSampler(self.sampler)
         if self.speculative_config is not None and self.speculative_config.method == "ngram":
             # 310P ngram requires decode-only graph shapes to be built with q_len=1.
             # Keep dispatcher's internal query_len in sync to avoid key-init assert.