remove rms_norm_opt and fused_add_rms_norm_opt

CMakeLists.txt

@@ -296,7 +296,7 @@ set(VLLM_EXT_SRC
   "csrc/layernorm_kernels.cu"
   "csrc/opt/transpose_kernels.cu"
   "csrc/opt/activation_kernels_opt.cu"
-  "csrc/opt/layernorm_kernels_opt.cu"
+  # "csrc/opt/layernorm_kernels_opt.cu"
   "csrc/fused_qknorm_rope_kernel.cu"
   # "csrc/layernorm_quant_kernels.cu"
   "csrc/sampler.cu"

csrc/ops.h

@@ -97,11 +97,11 @@ void fused_qk_norm_rope(torch::Tensor& qkv, int64_t num_heads_q,
                         torch::Tensor& k_weight, torch::Tensor& cos_sin_cache,
                         bool is_neox, torch::Tensor& position_ids);
 
-void rms_norm_opt(torch::Tensor& out, torch::Tensor& input, torch::Tensor& weight,
-              double epsilon);
+// void rms_norm_opt(torch::Tensor& out, torch::Tensor& input, torch::Tensor& weight,
+//               double epsilon);
 
-void fused_add_rms_norm_opt(torch::Tensor& input, torch::Tensor& residual,
-                        torch::Tensor& weight, double epsilon);
+// void fused_add_rms_norm_opt(torch::Tensor& input, torch::Tensor& residual,
+//                         torch::Tensor& weight, double epsilon);
 
 void apply_repetition_penalties_(torch::Tensor& logits,
                                  const torch::Tensor& prompt_mask,

csrc/torch_bindings.cpp

@@ -203,16 +203,16 @@ TORCH_LIBRARY_EXPAND(TORCH_EXTENSION_NAME, ops) {
 
   // Layernorm-quant
   // Apply Root Mean Square (RMS) Normalization to the input tensor.
-  ops.def(
-      "rms_norm_opt(Tensor! out, Tensor input, Tensor weight, float epsilon) -> "
-      "()");
-  ops.impl("rms_norm_opt", torch::kCUDA, &rms_norm_opt);
+//   ops.def(
+//       "rms_norm_opt(Tensor! out, Tensor input, Tensor weight, float epsilon) -> "
+//       "()");
+//   ops.impl("rms_norm_opt", torch::kCUDA, &rms_norm_opt);
 
   // In-place fused Add and RMS Normalization. (opt)
-  ops.def(
-      "fused_add_rms_norm_opt(Tensor! input, Tensor! residual, Tensor weight, "
-      "float epsilon) -> ()");
-  ops.impl("fused_add_rms_norm_opt", torch::kCUDA, &fused_add_rms_norm_opt);
+//   ops.def(
+//       "fused_add_rms_norm_opt(Tensor! input, Tensor! residual, Tensor weight, "
+//       "float epsilon) -> ()");
+//   ops.impl("fused_add_rms_norm_opt", torch::kCUDA, &fused_add_rms_norm_opt);
 
   // Layernorm-quant
   // Apply Root Mean Square (RMS) Normalization to the input tensor.

pyproject.toml

@@ -6,7 +6,7 @@ requires = [
     "packaging>=24.2",
     "setuptools>=77.0.3,<81.0.0",
     "setuptools-scm>=8.0",
-    "torch == 2.9.0",
+    "torch >= 2.7.1",
     "wheel",
     "jinja2",
 ]

requirements/rocm.txt

@@ -25,7 +25,7 @@ quart
 fastrlock==0.8.3
 cupy==12.3.0
 
-torch == 2.7.1
+torch >= 2.7.1
 triton == 3.1
 flash_attn == 2.6.1
 flash_mla == 1.0.0

vllm/_custom_ops.py

@@ -349,14 +349,14 @@ def fused_add_rms_norm(
     
 
 # layer norm ops (opt)
-def rms_norm_opt(out: torch.Tensor, input: torch.Tensor, weight: torch.Tensor,
-             epsilon: float) -> None:
-    torch.ops._C.rms_norm_opt(out, input, weight, epsilon)
+# def rms_norm_opt(out: torch.Tensor, input: torch.Tensor, weight: torch.Tensor,
+#              epsilon: float) -> None:
+#     torch.ops._C.rms_norm_opt(out, input, weight, epsilon)
 
 
-def fused_add_rms_norm_opt(input: torch.Tensor, residual: torch.Tensor,
-                       weight: torch.Tensor, epsilon: float) -> None:
-    torch.ops._C.fused_add_rms_norm_opt(input, residual, weight, epsilon)
+# def fused_add_rms_norm_opt(input: torch.Tensor, residual: torch.Tensor,
+#                        weight: torch.Tensor, epsilon: float) -> None:
+#     torch.ops._C.fused_add_rms_norm_opt(input, residual, weight, epsilon)
 
 
 def fused_qk_norm_rope(