[Feat] Add custom Embedding tensor model parallel (#2616)

Similar to #2309 , this PR introduces Embedding tensor model parallel to
achieve decreasing of memory consumption. It support both eager mode and
graph mode.

And this PR refactor module tensor parallel configurations supported in
#2309, #2167, #2120, merge all config into `finegrained_tp_config` in
`additional_config`, including:
`lmhead_tensor_parallel_size`
`oproj_tensor_parallel_size`
`embedding_tensor_parallel_size`
`mlp_tensor_parallel_size`

- vLLM version: v0.12.0
- vLLM main:
ad32e3e19c

---------

Signed-off-by: zzhx1 <zzh_201018@outlook.com>
Signed-off-by: zzhxx <zhangzihang23@mails.ucas.ac.cn>
Co-authored-by: zzhx1 <zzh_201018@outlook.com>
Co-authored-by: chenxiao <Jaychou1620@Gmail.com>
Co-authored-by: zzhxx <zhangzihang23@mails.ucas.ac.cn>
Co-authored-by: Jade Zheng <zheng.shoujian@outlook.com>

tests/ut/ops/test_vocab_parallel_embedding.py

@@ -14,11 +14,12 @@
 # Adapted from vllm/tests/lora/test_layers.py
 
 import unittest
+from unittest import mock
 from unittest.mock import MagicMock, patch
 
 import torch
 
-from vllm_ascend.ascend_config import init_ascend_config
+from vllm_ascend.distributed import parallel_state
 from vllm_ascend.ops.vocab_parallel_embedding import (
     AscendLogitsProcessor, AscendParallelLMHead, AscendVocabParallelEmbedding)
 
@@ -32,9 +33,33 @@ class TestCustomVocabParallelEmbedding(unittest.TestCase):
         self.embedding_dim = 10
         self.org_num_embeddings = 40
         self.padding_size = 8
+
+        self.mock_group = mock.MagicMock()
+        self.mock_group.world_size = 2
+        self.mock_group.rank_in_group = 0
+
+        parallel_state._MLP_TP = self.mock_group
+        parallel_state._OTP = self.mock_group
+
         mock_vllm_config = MagicMock()
         mock_vllm_config.additional_config = {}
-        init_ascend_config(mock_vllm_config)
+        self.mock_ascend_config = MagicMock()
+        self.mock_ascend_config.finegrained_tp_config.lmhead_tensor_parallel_size = 2
+        self.mock_ascend_config.finegrained_tp_config.embedding_tensor_parallel_size = 2
+
+        self.patches = [
+            patch("vllm_ascend.utils.get_ascend_config",
+                  return_value=self.mock_ascend_config),
+            patch("vllm_ascend.distributed.parallel_state.get_lmhead_tp_group",
+                  return_value=self.mock_group),
+            patch(
+                "vllm.distributed.parallel_state.get_tp_group",
+                return_value=self.mock_group,
+            ),
+        ]
+
+        for p in self.patches:
+            p.start()
 
     def _create_layer(self):
         # Patch methods and dependencies for VocabParallelEmbedding