layernorm_ref() — pytorch Function Reference

Architecture documentation for the layernorm_ref() function in kernels.py from the pytorch codebase.

Function python DynamoBenchmarks Configuration calls 1 called by 2

Entity Profile

DynamoBenchmarks→ Configuration→ layernorm_ref() — pytorch Function Reference

Dependency Diagram

graph TD
  7f08e9ae_5ab7_6b54_e9d5_a34f8e037133["layernorm_ref()"]
  e5a28a86_38cc_d5f9_6155_e07365660a34["layernorm_ref()"]
  e5a28a86_38cc_d5f9_6155_e07365660a34 -->|calls| 7f08e9ae_5ab7_6b54_e9d5_a34f8e037133
  fc3c62fa_4ffc_e8d4_5ee1_c74572a1d012["eager()"]
  fc3c62fa_4ffc_e8d4_5ee1_c74572a1d012 -->|calls| 7f08e9ae_5ab7_6b54_e9d5_a34f8e037133
  e5a28a86_38cc_d5f9_6155_e07365660a34["layernorm_ref()"]
  7f08e9ae_5ab7_6b54_e9d5_a34f8e037133 -->|calls| e5a28a86_38cc_d5f9_6155_e07365660a34
  style 7f08e9ae_5ab7_6b54_e9d5_a34f8e037133 fill:#6366f1,stroke:#818cf8,color:#fff

Relationship Graph

Source Code

benchmarks/dynamo/genai_layers/kernels.py lines 676–678

    def layernorm_ref(self, x: torch.Tensor, w: torch.Tensor, eps: float = 1e-6):
        x_f32 = x.float()
        return F.layer_norm(x_f32, w.shape, w, None, eps).to(x.dtype)