[Perf] Remove D2H operations to imporve performance (#4063)

### What this PR does / why we need it?
Replace masked in-place assignment with a device-side torch.where so
selection stays on-device, allowing subsequent device ops to be enqueued
earlier and removing an implicit D2H sync, reducing latency by several
hundreds μs on Ascend.

### Does this PR introduce _any_ user-facing change?
None.

### How was this patch tested?
None.
- vLLM version: v0.11.0
- vLLM main:
83f478bb19

Signed-off-by: Yizhou Liu <liu_yizhou@outlook.com>

vllm_ascend/sample/rejection_sampler.py

@@ -320,7 +320,8 @@ def rejection_greedy_sample_spec_len_1_pytorch(
     accept_req_mask = draft_token_ids == target_argmax
     output_token_ids[:, 0] = target_argmax
     bonus_token_ids = bonus_token_ids.squeeze(1)
-    output_token_ids[accept_req_mask, 1] = bonus_token_ids[accept_req_mask]
+    output_token_ids[:, 1] = torch.where(accept_req_mask, bonus_token_ids,
+                                         output_token_ids[:, 1])
 
 
 def rejection_greedy_sample_pytorch(