[EPLB][Bugfix] Dispatch Allgather use log2phy if enable eplb (#5933)

### What this PR does / why we need it?
1. Move the logic of expert mapping forward to prevent shotgun changes
2. Disable the update of expert map.

### How was this patch tested?
a2
| dataset | version | metric | mode | vllm-api-general-chat |
|----- | ----- | ----- | ----- | -----|
| GPQA_diamond | 53064e | accuracy | gen | 73.23 |

a3
| dataset | version | metric | mode | vllm-api-general-chat |
|----- | ----- | ----- | ----- | -----|
| aime2024 | 604a78 | accuracy | gen | 83.33 |


- vLLM version: v0.13.0
- vLLM main:
11b6af5280

Signed-off-by: shenchuxiaofugui <1311027364@qq.com>

tests/ut/ops/test_token_dispatcher.py

@@ -472,23 +472,3 @@ class TestTokenDispatcherWithAll2AllV(TestBase):
         self.assertIsNotNone(result.dynamic_scale)
         self.assertEqual(result.group_list_type, 1)
 
-    def test_token_dispatch_with_log2phy(self):
-        hidden_states = torch.randn(8, 16)
-        topk_weights = torch.rand(8, 4)
-        topk_ids = torch.randint(0, 4, (8, 2)).long()
-        expert_map = torch.tensor([0, 1, 2, 3])
-        log2phy = torch.tensor([1, 0, 3, 2])
-
-        self.dispatcher.expert_ids_per_ep_rank = torch.tensor(
-            [0, 1], dtype=torch.int32)
-        self.dispatcher.local_expert_indices = [0, 1]
-
-        result = self.dispatcher.token_dispatch(hidden_states=hidden_states,
-                                                topk_weights=topk_weights,
-                                                topk_ids=topk_ids,
-                                                expert_map=expert_map,
-                                                log2phy=log2phy)
-
-        self.assertIsNotNone(result.hidden_states)
-        self.assertIsNotNone(result.group_list)
-        self.assertEqual(result.group_list_type, 1)