add m_grouped_bf16_gemm_nn_contiguous, del paddle.batch_gemm (#7527)

fastdeploy/model_executor/layers/moe/fused_moe_cutlass_backend.py

@@ -53,6 +53,12 @@ from fastdeploy.model_executor.utils import (
 )
 
 
+def m_grouped_bf16_gemm_nn_contiguous(x, y, expert_idx_per_token):
+    out = paddle.empty([x.shape[0], y.shape[-1]], dtype=x.dtype)
+    paddlefleet_ops.deep_gemm.m_grouped_bf16_gemm_nn_contiguous(x, y, out, expert_idx_per_token)
+    return out
+
+
 class CutlassMoEMethod(UnquantizedFusedMoEMethod):
     """
     Use Cutlass Group Gemm to compute Fused MoE.
@@ -156,7 +162,8 @@ class CutlassMoEMethod(UnquantizedFusedMoEMethod):
             if fastdeploy.envs.FD_USE_PHI_MOE_PERMUTE and self.moe_quant_type == "w16a16":
                 # --- moe_permute / moe_unpermute path ---
                 recv_topk_idx_i32 = recv_topk_idx.astype(paddle.int32)
-                (permute_input, permute_indices_per_token, dst_weights, _scale_out) = paddle.nn.functional.moe_permute(
+                (permute_input, permute_indices_per_token, dst_weights, _scale_out, m_indices) = (
+                    paddle.nn.functional.moe_permute(
                         hidden_states=recv_x,
                         scale=None,
                         expert_routemap_topk=recv_topk_idx_i32,
@@ -165,17 +172,31 @@ class CutlassMoEMethod(UnquantizedFusedMoEMethod):
                         tokens_per_expert=[],
                         padding_alignment=128,
                         override_buffer_size=token_all_num,
+                        return_expert_indices=True,
+                    )
                 )
 
+                if paddlefleet_ops is not None:
+                    out = m_grouped_bf16_gemm_nn_contiguous(
+                        permute_input, getattr(layer, self.added_weight_attrs[0]), m_indices
+                    )
+                else:
                     out = paddle.incubate.nn.functional.batched_gemm(
                         permute_input,
                         getattr(layer, self.added_weight_attrs[0]),
                         recv_num_tokens_per_expert_list,
                     )
+
                 if fastdeploy.envs.FD_MOE_PROB_IN_ADVANCE:
                     out = paddlefleet_ops.fused_swiglu_scale(out, dst_weights)
                 else:
                     out = paddle.incubate.nn.functional.swiglu(out)
+
+                if paddlefleet_ops is not None:
+                    ffn_out = m_grouped_bf16_gemm_nn_contiguous(
+                        out, getattr(layer, self.added_weight_attrs[1]), m_indices
+                    )
+                else:
                     ffn_out = paddle.incubate.nn.functional.batched_gemm(
                         out,
                         getattr(layer, self.added_weight_attrs[1]),