Optimize GeGLU layer in Gemma (vllm-project#2975)

Author: WoosukKwon

csrc/activation_kernels.cu

@@ -2,52 +2,75 @@
 #include <torch/extension.h>
 #include <c10/cuda/CUDAGuard.h>
 
+#include <cmath>
+
 #include "cuda_compat.h"
 #include "dispatch_utils.h"
 
 namespace vllm {
 
-template<typename T>
-__device__ __forceinline__ T silu(const T& x) {
-  // x * sigmoid(x)
-  return (T) (((float) x) / (1.0f + expf((float) -x)));
-}
-
-template<typename scalar_t>
-__global__ void silu_and_mul_kernel(
+// Activation and gating kernel template.
+template<typename scalar_t, scalar_t (*ACT_FN)(const scalar_t&)>
+__global__ void act_and_mul_kernel(
   scalar_t* __restrict__ out,               // [..., d]
   const scalar_t* __restrict__ input,       // [..., 2, d]
   const int d) {
   const int64_t token_idx = blockIdx.x;
   for (int64_t idx = threadIdx.x; idx < d; idx += blockDim.x) {
     const scalar_t x = VLLM_LDG(&input[token_idx * 2 * d + idx]);
     const scalar_t y = VLLM_LDG(&input[token_idx * 2 * d + d + idx]);
-    out[token_idx * d + idx] = silu(x) * y;
+    out[token_idx * d + idx] = ACT_FN(x) * y;
   }
 }
 
+template<typename T>
+__device__ __forceinline__ T silu_kernel(const T& x) {
+  // x * sigmoid(x)
+  return (T) (((float) x) / (1.0f + expf((float) -x)));
+}
+
+template<typename T>
+__device__ __forceinline__ T gelu_kernel(const T& x) {
+  // Equivalent to PyTorch GELU with 'none' approximation.
+  // Refer to:
+  // https://github.com/pytorch/pytorch/blob/8ac9b20d4b090c213799e81acf48a55ea8d437d6/aten/src/ATen/native/cuda/ActivationGeluKernel.cu#L38
+  const float f = (float) x;
+  constexpr float ALPHA = M_SQRT1_2;
+  return (T) (f * 0.5f * (1.0f + ::erf(f * ALPHA)));
+}
+
 } // namespace vllm
 
+// Launch activation and gating kernel.
+#define LAUNCH_ACTIVATION_GATE_KERNEL(KERNEL)                                             \
+  int d = input.size(-1) / 2;                                                             \
+  int64_t num_tokens = input.numel() / input.size(-1);                                    \
+  dim3 grid(num_tokens);                                                                  \
+  dim3 block(std::min(d, 1024));                                                          \
+  const at::cuda::OptionalCUDAGuard device_guard(device_of(input));                       \
+  const cudaStream_t stream = at::cuda::getCurrentCUDAStream();                           \
+  VLLM_DISPATCH_FLOATING_TYPES(                                                           \
+    input.scalar_type(),                                                                  \
+    "act_and_mul_kernel",                                                                 \
+    [&] {                                                                                 \
+      vllm::act_and_mul_kernel<scalar_t, KERNEL<scalar_t>><<<grid, block, 0, stream>>>(   \
+        out.data_ptr<scalar_t>(),                                                         \
+        input.data_ptr<scalar_t>(),                                                       \
+        d);                                                                               \
+    });
+
 void silu_and_mul(
   torch::Tensor& out,      // [..., d]
   torch::Tensor& input)    // [..., 2 * d]
 {
-  int64_t num_tokens = input.numel() / input.size(-1);
-  int d = input.size(-1) / 2;
-
-  dim3 grid(num_tokens);
-  dim3 block(std::min(d, 1024));
-  const at::cuda::OptionalCUDAGuard device_guard(device_of(input));
-  const cudaStream_t stream = at::cuda::getCurrentCUDAStream();
-  VLLM_DISPATCH_FLOATING_TYPES(
-    input.scalar_type(),
-    "silu_and_mul_kernel",
-    [&] {
-      vllm::silu_and_mul_kernel<scalar_t><<<grid, block, 0, stream>>>(
-        out.data_ptr<scalar_t>(),
-        input.data_ptr<scalar_t>(),
-        d);
-    });
+  LAUNCH_ACTIVATION_GATE_KERNEL(vllm::silu_kernel);
+}
+
+void gelu_and_mul(
+  torch::Tensor& out,      // [..., d]
+  torch::Tensor& input)    // [..., 2 * d]
+{
+  LAUNCH_ACTIVATION_GATE_KERNEL(vllm::gelu_kernel);
 }
 
 namespace vllm {

csrc/ops.h

@@ -57,6 +57,10 @@ void silu_and_mul(
   torch::Tensor& out,
   torch::Tensor& input);
 
+void gelu_and_mul(
+  torch::Tensor& out,
+  torch::Tensor& input);
+
 void gelu_new(
   torch::Tensor& out,
   torch::Tensor& input);

csrc/pybind.cpp

@@ -22,6 +22,10 @@ PYBIND11_MODULE(TORCH_EXTENSION_NAME, m) {
     "silu_and_mul",
     &silu_and_mul,
     "Activation function used in SwiGLU.");
+  ops.def(
+    "gelu_and_mul",
+    &gelu_and_mul,
+    "Activation function used in GeGLU.");
   ops.def(
     "gelu_new",
     &gelu_new,

tests/kernels/test_activation.py

@@ -1,7 +1,10 @@
+from typing import Type
+
 import pytest
 import torch
 
-from vllm.model_executor.layers.activation import FastGELU, NewGELU, SiluAndMul
+from vllm.model_executor.layers.activation import (FastGELU, GeluAndMul,
+                                                   NewGELU, SiluAndMul)
 from allclose_default import get_default_atol, get_default_rtol
 
 DTYPES = [torch.half, torch.bfloat16, torch.float]
@@ -13,13 +16,15 @@
 ]
 
 
+@pytest.mark.parametrize("activation", [SiluAndMul, GeluAndMul])
 @pytest.mark.parametrize("num_tokens", NUM_TOKENS)
 @pytest.mark.parametrize("d", D)
 @pytest.mark.parametrize("dtype", DTYPES)
 @pytest.mark.parametrize("seed", SEEDS)
 @pytest.mark.parametrize("device", CUDA_DEVICES)
 @torch.inference_mode()
-def test_silu_and_mul(
+def test_act_and_mul(
+    activation: Type[torch.nn.Module],
     num_tokens: int,
     d: int,
     dtype: torch.dtype,
@@ -31,48 +36,23 @@ def test_silu_and_mul(
         torch.cuda.manual_seed(seed)
     torch.set_default_device(device)
     x = torch.randn(num_tokens, 2 * d, dtype=dtype)
-    layer = SiluAndMul()
+    layer = activation()
     out = layer(x)
     ref_out = layer._forward(x)
-    assert torch.allclose(out,
-                          ref_out,
-                          atol=get_default_atol(out),
-                          rtol=get_default_rtol(out))
+    # The SiLU and GELU implementations are equivalent to the native PyTorch
+    # implementations, so we can do exact comparison.
+    assert torch.allclose(out, ref_out, atol=0.0, rtol=0.0)
 
 
+@pytest.mark.parametrize("activation", [FastGELU, NewGELU])
 @pytest.mark.parametrize("num_tokens", NUM_TOKENS)
 @pytest.mark.parametrize("d", D)
 @pytest.mark.parametrize("dtype", DTYPES)
 @pytest.mark.parametrize("seed", SEEDS)
 @pytest.mark.parametrize("device", CUDA_DEVICES)
 @torch.inference_mode()
-def test_gelu_new(
-    num_tokens: int,
-    d: int,
-    dtype: torch.dtype,
-    seed: int,
-    device: str,
-) -> None:
-    torch.random.manual_seed(seed)
-    if torch.cuda.is_available():
-        torch.cuda.manual_seed(seed)
-    torch.set_default_device(device)
-    x = torch.randn(num_tokens, d, dtype=dtype)
-    layer = NewGELU()
-    out = layer(x)
-    ref_out = layer._forward(x)
-    assert torch.allclose(out,
-                          ref_out,
-                          atol=get_default_atol(out),
-                          rtol=get_default_rtol(out))
-
-
-@pytest.mark.parametrize("num_tokens", NUM_TOKENS)
-@pytest.mark.parametrize("d", D)
-@pytest.mark.parametrize("dtype", DTYPES)
-@pytest.mark.parametrize("seed", SEEDS)
-@pytest.mark.parametrize("device", CUDA_DEVICES)
-def test_gelu_fast(
+def test_activation(
+    activation: Type[torch.nn.Module],
     num_tokens: int,
     d: int,
     dtype: torch.dtype,
@@ -84,7 +64,7 @@ def test_gelu_fast(
         torch.cuda.manual_seed(seed)
     torch.set_default_device(device)
     x = torch.randn(num_tokens, d, dtype=dtype)
-    layer = FastGELU()
+    layer = activation()
     out = layer(x)
     ref_out = layer._forward(x)
     assert torch.allclose(out,

vllm/model_executor/layers/activation.py

@@ -37,6 +37,29 @@ def forward(self, x: torch.Tensor) -> torch.Tensor:
         return out
 
 
+class GeluAndMul(nn.Module):
+    """An activation function for GeGLU.
+
+    The function computes x -> GELU(x[:d]) * x[d:] where d = x.shape[-1] // 2.
+
+    Shapes:
+        x: (batch_size, seq_len, 2 * d) or (num_tokens, 2 * d)
+        return: (batch_size, seq_len, d) or (num_tokens, d)
+    """
+
+    def _forward(self, x: torch.Tensor) -> torch.Tensor:
+        """PyTorch-native implementation equivalent to forward()."""
+        d = x.shape[-1] // 2
+        return F.gelu(x[..., :d]) * x[..., d:]
+
+    def forward(self, x: torch.Tensor) -> torch.Tensor:
+        d = x.shape[-1] // 2
+        output_shape = (x.shape[:-1] + (d, ))
+        out = torch.empty(output_shape, dtype=x.dtype, device=x.device)
+        ops.gelu_and_mul(out, x)
+        return out
+
+
 class NewGELU(nn.Module):
 
     def _forward(self, x: torch.Tensor) -> torch.Tensor:

vllm/model_executor/models/gemma.py

@@ -21,10 +21,11 @@
 from transformers import GemmaConfig
 
 from vllm.model_executor.input_metadata import InputMetadata
+from vllm.model_executor.layers.activation import GeluAndMul
 from vllm.model_executor.layers.attention import PagedAttention
 from vllm.model_executor.layers.layernorm import RMSNorm
-from vllm.model_executor.layers.linear import (ColumnParallelLinear,
-                                               LinearMethodBase,
+from vllm.model_executor.layers.linear import (LinearMethodBase,
+                                               MergedColumnParallelLinear,
                                                QKVParallelLinear,
                                                RowParallelLinear)
 from vllm.model_executor.layers.rotary_embedding import get_rope
@@ -50,27 +51,21 @@ def __init__(
         linear_method: Optional[LinearMethodBase] = None,
     ) -> None:
         super().__init__()
-        self.gate_proj = ColumnParallelLinear(hidden_size,
-                                              intermediate_size,
-                                              bias=False,
-                                              linear_method=linear_method)
-        self.up_proj = ColumnParallelLinear(hidden_size,
-                                            intermediate_size,
-                                            bias=False,
-                                            linear_method=linear_method)
+        self.gate_up_proj = MergedColumnParallelLinear(
+            hidden_size, [intermediate_size] * 2,
+            bias=False,
+            linear_method=linear_method)
         self.down_proj = RowParallelLinear(intermediate_size,
                                            hidden_size,
                                            bias=False,
                                            linear_method=linear_method)
-        self.act_fn = nn.GELU()
+        self.act_fn = GeluAndMul()
 
     def forward(self, x):
-        gate, _ = self.gate_proj(x)
-        gate = self.act_fn(gate)
-        up, _ = self.up_proj(x)
-        fuse = gate * up
-        outputs, _ = self.down_proj(fuse)
-        return outputs
+        gate_up, _ = self.gate_up_proj(x)
+        x = self.act_fn(gate_up)
+        x, _ = self.down_proj(x)
+        return x
 
 
 class GemmaAttention(nn.Module):
@@ -294,6 +289,8 @@ def load_weights(self,
             ("qkv_proj", "q_proj", "q"),
             ("qkv_proj", "k_proj", "k"),
             ("qkv_proj", "v_proj", "v"),
+            ("gate_up_proj", "gate_proj", 0),
+            ("gate_up_proj", "up_proj", 1),
         ]
         params_dict = dict(self.named_parameters())
         loaded_params = set()