activation_kernels_opt.cu 8.17 KB
Newer Older
zhuwenwen's avatar
zhuwenwen committed
1
2
3
4
5
6
7
8
9
10
11
12
#include <ATen/cuda/CUDAContext.h>
#include <torch/all.h>
#include <c10/cuda/CUDAGuard.h>
#include <ATen/native/cuda/MemoryAccess.cuh>

#include <cmath>

#include "cuda_compat.h"
#include "../dispatch_utils.h"

namespace vllm {

zhuwenwen's avatar
zhuwenwen committed
13
14
15
16
17
18
template <typename scalar_t, scalar_t (*ACT_FN)(const scalar_t&),
          bool act_first>
__device__ __forceinline__ scalar_t compute(const scalar_t& x,
                                            const scalar_t& y) {
  return act_first ? ACT_FN(x) * y : x * ACT_FN(y);
}
zhuwenwen's avatar
zhuwenwen committed
19
// Activation and gating kernel template.
zhuwenwen's avatar
zhuwenwen committed
20
21
template <typename scalar_t, scalar_t (*ACT_FN)(const scalar_t&),
          bool act_first>
zhuwenwen's avatar
zhuwenwen committed
22
23
24
25
26
27
28
29
__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]);
zhuwenwen's avatar
zhuwenwen committed
30
    out[token_idx * d + idx] = compute<scalar_t, ACT_FN, act_first>(x, y);
zhuwenwen's avatar
zhuwenwen committed
31
32
33
  }
}

zhuwenwen's avatar
zhuwenwen committed
34
35
template <typename scalar_t, scalar_t (*ACT_FN)(const scalar_t&), int VEC,
          bool act_first>
zhuwenwen's avatar
zhuwenwen committed
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
__global__ void act_and_mul_kernel_opt1(
    scalar_t* __restrict__ out,          // [..., d]
    const scalar_t* __restrict__ input,  // [..., 2, d]
    const int d) {
  using VecType = at::native::memory::aligned_vector<scalar_t, VEC>;
  const int64_t token_idx= blockIdx.x;
  int idx = threadIdx.x * VEC;
  if (idx < d) {
    const int64_t x_index = token_idx * 2 * d + idx;
    const int64_t y_index = token_idx * d + idx;
    VecType* x1 = (VecType*)(input + x_index);
    VecType* x2 = (VecType*)(input + x_index + d);
    VecType* y = (VecType*)(out + y_index);
    scalar_t r_x1[VEC];
    scalar_t r_x2[VEC];
    scalar_t r_y[VEC];
    *(VecType*)r_x1 = *x1;
    *(VecType*)r_x2 = *x2;
#pragma unroll
    for (int i = 0; i < VEC; i++) {
      r_y[i] = ACT_FN(r_x1[i]) * r_x2[i];
    }
    *y = *(VecType*)r_y;
  }
}

zhuwenwen's avatar
zhuwenwen committed
62
63
template <typename scalar_t, scalar_t (*ACT_FN)(const scalar_t&), int VEC,
          bool act_first>
zhuwenwen's avatar
zhuwenwen committed
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
__global__ void act_and_mul_kernel_opt2(
    scalar_t* __restrict__ out,          // [..., d]
    const scalar_t* __restrict__ input,  // [..., 2, d]
    const int d) {
  using VecType = at::native::memory::aligned_vector<scalar_t, VEC>;
  const int64_t token_idx = blockIdx.x;
  int idx = threadIdx.x * VEC;
  for (; idx < d; idx += blockDim.x * VEC) {
    const int64_t x_index = token_idx * 2 * d + idx;
    const int64_t y_index = token_idx * d + idx;
    VecType* x1 = (VecType*)(input + x_index);
    VecType* x2 = (VecType*)(input + x_index + d);
    VecType* y = (VecType*)(out + y_index);
    scalar_t r_x1[VEC];
    scalar_t r_x2[VEC];
    scalar_t r_y[VEC];
    *(VecType*)r_x1 = *x1;
    *(VecType*)r_x2 = *x2;
#pragma unroll
    for (int i = 0; i < VEC; i++) {
      r_y[i] = ACT_FN(r_x1[i]) * r_x2[i];
    }
    *y = *(VecType*)r_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#L36-L38
  const float f = (float)x;
  constexpr float ALPHA = M_SQRT1_2;
  return (T)(f * 0.5f * (1.0f + ::erf(f * ALPHA)));
}

template <typename T>
__device__ __forceinline__ T gelu_tanh_kernel(const T& x) {
  // Equivalent to PyTorch GELU with 'tanh' approximation.
  // Refer to:
  // https://github.com/pytorch/pytorch/blob/8ac9b20d4b090c213799e81acf48a55ea8d437d6/aten/src/ATen/native/cuda/ActivationGeluKernel.cu#L25-L30
  const float f = (float)x;
  constexpr float BETA = M_SQRT2 * M_2_SQRTPI * 0.5f;
  constexpr float KAPPA = 0.044715;
  float x_cube = f * f * f;
  float inner = BETA * (f + KAPPA * x_cube);
  return (T)(0.5f * f * (1.0f + ::tanhf(inner)));
}
zhuwenwen's avatar
zhuwenwen committed
118
    
zhuwenwen's avatar
zhuwenwen committed
119
120
}  // namespace vllm

zhuwenwen's avatar
zhuwenwen committed
121

zhuwenwen's avatar
zhuwenwen committed
122
123
124
125
// Launch activation and gating kernel.
// Use ACT_FIRST (bool) indicating whether to apply the activation function
// first.
#define LAUNCH_ACTIVATION_GATE_KERNEL(KERNEL, ACT_FIRST)                                  \
zhuwenwen's avatar
zhuwenwen committed
126
127
128
129
  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));                                               \
zhuwenwen's avatar
zhuwenwen committed
130
131
  if (num_tokens == 0) {                                                       \
    return;                                                                    \
zhuwenwen's avatar
zhuwenwen committed
132
  }                                                                            \
zhuwenwen's avatar
zhuwenwen committed
133
134
135
136
137
138
  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", [&] {                         \
        if (0 == d % 8 && d <= 16384) {                                        \
          if (d <= 512) {                                                      \
zhuwenwen's avatar
zhuwenwen committed
139
            vllm::act_and_mul_kernel_opt1<scalar_t, KERNEL<scalar_t>, 2, ACT_FIRST> \
zhuwenwen's avatar
zhuwenwen committed
140
141
142
                <<<grid, 256, 0, stream>>>(out.data_ptr<scalar_t>(),           \
                                           input.data_ptr<scalar_t>(), d);     \
          } else if (d <= 1024) {                                              \
zhuwenwen's avatar
zhuwenwen committed
143
            vllm::act_and_mul_kernel_opt1<scalar_t, KERNEL<scalar_t>, 8, ACT_FIRST> \
zhuwenwen's avatar
zhuwenwen committed
144
145
146
                <<<grid, 128, 0, stream>>>(out.data_ptr<scalar_t>(),           \
                                           input.data_ptr<scalar_t>(), d);     \
          } else if (d <= 2048) {                                              \
zhuwenwen's avatar
zhuwenwen committed
147
            vllm::act_and_mul_kernel_opt1<scalar_t, KERNEL<scalar_t>, 8, ACT_FIRST> \
zhuwenwen's avatar
zhuwenwen committed
148
149
150
                <<<grid, 256, 0, stream>>>(out.data_ptr<scalar_t>(),           \
                                           input.data_ptr<scalar_t>(), d);     \
          } else if (d <= 4096) {                                              \
zhuwenwen's avatar
zhuwenwen committed
151
            vllm::act_and_mul_kernel_opt1<scalar_t, KERNEL<scalar_t>, 8, ACT_FIRST> \
zhuwenwen's avatar
zhuwenwen committed
152
153
154
                <<<grid, 512, 0, stream>>>(out.data_ptr<scalar_t>(),           \
                                           input.data_ptr<scalar_t>(), d);     \
          } else {                                                             \
zhuwenwen's avatar
zhuwenwen committed
155
            vllm::act_and_mul_kernel_opt2<scalar_t, KERNEL<scalar_t>, 8, ACT_FIRST> \
zhuwenwen's avatar
zhuwenwen committed
156
157
158
159
                <<<grid, 1024, 0, stream>>>(out.data_ptr<scalar_t>(),          \
                                            input.data_ptr<scalar_t>(), d);    \
          }                                                                    \
        } else {                                                               \
zhuwenwen's avatar
zhuwenwen committed
160
              vllm::act_and_mul_kernel<scalar_t, KERNEL<scalar_t>, ACT_FIRST>             \
zhuwenwen's avatar
zhuwenwen committed
161
162
163
164
165
166
167
168
                  <<<grid, block, 0, stream>>>(out.data_ptr<scalar_t>(),       \
                                              input.data_ptr<scalar_t>(), d);  \
        }                                                                      \
      });

void silu_and_mul_opt(torch::Tensor& out,    // [..., d]
                  torch::Tensor& input)  // [..., 2 * d]
{
zhuwenwen's avatar
zhuwenwen committed
169
  LAUNCH_ACTIVATION_GATE_KERNEL(vllm::silu_kernel, true);
zhuwenwen's avatar
zhuwenwen committed
170
171
}

zhuwenwen's avatar
zhuwenwen committed
172
173
174
175
176
177
178
179
// void mul_and_silu_opt(torch::Tensor& out,    // [..., d]
//                   torch::Tensor& input)  // [..., 2 * d]
// {
//   // The difference between mul_and_silu and silu_and_mul is that mul_and_silu
//   // applies the silu to the latter half of the input.
//   LAUNCH_ACTIVATION_GATE_KERNEL(vllm::silu_kernel, false);
// }

zhuwenwen's avatar
zhuwenwen committed
180
181
182
void gelu_and_mul_opt(torch::Tensor& out,    // [..., d]
                  torch::Tensor& input)  // [..., 2 * d]
{
zhuwenwen's avatar
zhuwenwen committed
183
  LAUNCH_ACTIVATION_GATE_KERNEL(vllm::gelu_kernel, true);
zhuwenwen's avatar
zhuwenwen committed
184
185
186
187
188
}

void gelu_tanh_and_mul_opt(torch::Tensor& out,    // [..., d]
                       torch::Tensor& input)  // [..., 2 * d]
{
zhuwenwen's avatar
zhuwenwen committed
189
  LAUNCH_ACTIVATION_GATE_KERNEL(vllm::gelu_tanh_kernel, true);
zhuwenwen's avatar
zhuwenwen committed
190
}