use apply_rotary_emb_torch on z100l&k100 (qwen2-vl)

vllm/model_executor/models/qwen2_vl.py

@@ -85,6 +85,7 @@ import re
 from vllm import _custom_ops as ops
 from vllm.model_executor.utils import pad_weight, gemm_bank_conf
 from vllm.platforms import current_platform
+from vllm.utils import SUPPORT_TC
 
 logger = init_logger(__name__)
 
@@ -246,7 +247,7 @@ def apply_rotary_pos_emb_vision(t: torch.Tensor,
     apply_rotary_emb = apply_rotary_emb_torch
     if current_platform.is_cuda():
         from vllm.vllm_flash_attn.layers.rotary import apply_rotary_emb
-    if current_platform.is_rocm():
+    if current_platform.is_rocm() and SUPPORT_TC:
         from flash_attn.layers.rotary import apply_rotary_emb
     output = apply_rotary_emb(t_, cos, sin).type_as(t)
     return output