[refact] unified soc_version code (#4359)

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

Currently, there are two paths to judge the chip type in code,
`get_ascend_soc_version` use `get_soc_version` api in torch_npu, and
`is_310p` `use _build_info.__soc_version__`, which generate when
install. We need to unify the two paths.

We need to unify these codes based on the following points:

1. We need to ensure consistency in chip type judgment between compiling
and running states;
2. In compiling state, we need chip type to complete op's compilation,
but in running state, we only need device
type(910B/910_93/310P/910_95/etc) to make code branch judgement;
3. In compiling state, torch_npu may not have been installed yet, so we
can't use torch_npu's api.

Based on the above points, we have made the following changes:

1. When user set env `SOC_VERSION`, use it; when not set, query
soc_version by `npu-smi`;
2. generate device_type based on soc_version when compiling, and write
`__device_type__` instead of `__soc_version__` in `_build_info.py`;
3. In running state, use `__device_type__` to judge code branch.

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

When not set env `SOC_VERSION`, it will not be `ASCEND910B1` by default,
we will query soc_version by `npu-smi`. And env `SOC_VERSION` must be in
the list `soc_to_device` in `setup.py`.

- vLLM version: v0.11.0
- vLLM main:
2918c1b49c

Signed-off-by: zzzzwwjj <1183291235@qq.com>

vllm_ascend/ops/fused_moe/fused_moe.py

@@ -43,9 +43,9 @@ from vllm_ascend.quantization.w4a8_dynamic import \
     AscendW4A8DynamicFusedMoEMethod
 from vllm_ascend.quantization.w8a8_dynamic import \
     AscendW8A8DynamicFusedMoEMethod
-from vllm_ascend.utils import (ACL_FORMAT_FRACTAL_NZ, enable_sp, is_310p,
-                               is_enable_nz, npu_stream_switch,
-                               shared_expert_dp_enabled,
+from vllm_ascend.utils import (ACL_FORMAT_FRACTAL_NZ, AscendDeviceType,
+                               enable_sp, get_ascend_device_type, is_enable_nz,
+                               npu_stream_switch, shared_expert_dp_enabled,
                                shared_experts_calculation_stream)
 
 
@@ -79,7 +79,8 @@ class AscendUnquantizedFusedMoEMethod(UnquantizedFusedMoEMethod):
             w2_data = self._maybe_pad_weight(layer.w2_weight.data)
             layer.w2_weight = torch.nn.Parameter(w2_data, requires_grad=False)
 
-        if not is_310p() and is_enable_nz():
+        if get_ascend_device_type() != AscendDeviceType._310P and is_enable_nz(
+        ):
             layer.w13_weight.data = torch_npu.npu_format_cast(
                 layer.w13_weight.data, ACL_FORMAT_FRACTAL_NZ)
             layer.w2_weight.data = torch_npu.npu_format_cast(

vllm_ascend/ops/fused_moe/moe_mlp.py

@@ -22,7 +22,8 @@ from torch.nn.functional import pad
 from vllm.forward_context import get_forward_context
 
 from vllm_ascend.ascend_forward_context import MoECommType
-from vllm_ascend.utils import dispose_tensor, is_310p
+from vllm_ascend.utils import (AscendDeviceType, dispose_tensor,
+                               get_ascend_device_type)
 
 
 def cumsum_group_list(group_list: torch.Tensor,
@@ -210,7 +211,7 @@ def unquant_apply_mlp(hidden_states: torch.Tensor,
         group_type=0,
         group_list=group_list,
     )[0]
-    if is_310p():
+    if get_ascend_device_type() == AscendDeviceType._310P:
         gate_up_out = torch_npu.npu_swiglu(gate_up_out.to(torch.float32)).to(
             torch.float16)
     else:

vllm_ascend/ops/fused_moe/token_dispatcher.py

@@ -30,7 +30,7 @@ from vllm.distributed.parallel_state import get_ep_group
 from vllm_ascend.distributed.parallel_state import get_mc2_group
 from vllm_ascend.ops.fused_moe.comm_utils import (
     async_all_to_all, gather_from_sequence_parallel_region)
-from vllm_ascend.utils import (AscendSocVersion, get_ascend_soc_version,
+from vllm_ascend.utils import (AscendDeviceType, get_ascend_device_type,
                                is_hierarchical_communication_enabled)
 
 
@@ -98,11 +98,11 @@ class TokenDispatcherWithMC2(MoETokenDispatcher):
         self.enable_dispatch_v2 = hasattr(torch_npu,
                                           "npu_moe_distribute_dispatch_v2")
         self.need_extra_args = (
-            get_ascend_soc_version() == AscendSocVersion.A3)
+            get_ascend_device_type() == AscendDeviceType._910_93)
 
         # NOTE: Currently, when in A3, we need to pass in some extra param into dispatch & combine
         self.a3_need_extra_args = \
-            get_ascend_soc_version() == AscendSocVersion.A3
+            get_ascend_device_type() == AscendDeviceType._910_93
         # NOTE: When in A2, setting the environment variables HCCL_INTRA_PCIE_ENABLE=1 and
         # HCCL_INTRA_ROCE_ENABLE=0 can reduce cross-machine communication traffic and significantly
         # improve communication performance.