2024-03-25 13:50:23 +01:00
|
|
|
#include "common.cuh"
|
|
|
|
|
|
2024-09-03 17:21:46 +02:00
|
|
|
#define CUDA_NEG_BLOCK_SIZE 256
|
2024-09-20 19:04:44 +03:00
|
|
|
#define CUDA_STEP_BLOCK_SIZE 256
|
2024-03-25 13:50:23 +01:00
|
|
|
#define CUDA_GELU_BLOCK_SIZE 256
|
|
|
|
|
#define CUDA_SILU_BLOCK_SIZE 256
|
2025-01-16 16:43:38 +01:00
|
|
|
#define CUDA_SILU_BACK_BLOCK_SIZE 256
|
2024-03-25 13:50:23 +01:00
|
|
|
#define CUDA_TANH_BLOCK_SIZE 256
|
|
|
|
|
#define CUDA_RELU_BLOCK_SIZE 256
|
2024-05-01 14:44:26 -07:00
|
|
|
#define CUDA_SIGMOID_BLOCK_SIZE 256
|
2024-03-25 13:50:23 +01:00
|
|
|
#define CUDA_HARDSIGMOID_BLOCK_SIZE 256
|
2024-09-22 10:29:12 +08:00
|
|
|
#define CUDA_EXP_BLOCK_SIZE 256
|
2024-03-25 13:50:23 +01:00
|
|
|
#define CUDA_HARDSWISH_BLOCK_SIZE 256
|
|
|
|
|
#define CUDA_SQR_BLOCK_SIZE 256
|
2024-06-16 15:23:04 -07:00
|
|
|
#define CUDA_SQRT_BLOCK_SIZE 256
|
2024-08-27 22:01:45 +03:00
|
|
|
#define CUDA_SIN_BLOCK_SIZE 256
|
|
|
|
|
#define CUDA_COS_BLOCK_SIZE 256
|
2025-06-29 11:04:10 +02:00
|
|
|
#define CUDA_GLU_BLOCK_SIZE 256
|
2024-03-25 13:50:23 +01:00
|
|
|
|
2025-02-28 12:34:39 +05:30
|
|
|
void ggml_cuda_op_abs(ggml_backend_cuda_context & ctx, ggml_tensor * dst);
|
|
|
|
|
|
|
|
|
|
void ggml_cuda_op_sgn(ggml_backend_cuda_context & ctx, ggml_tensor * dst);
|
|
|
|
|
|
2024-09-03 17:21:46 +02:00
|
|
|
void ggml_cuda_op_neg(ggml_backend_cuda_context & ctx, ggml_tensor * dst);
|
|
|
|
|
|
2024-09-20 19:04:44 +03:00
|
|
|
void ggml_cuda_op_step(ggml_backend_cuda_context & ctx, ggml_tensor * dst);
|
|
|
|
|
|
2024-03-25 13:50:23 +01:00
|
|
|
void ggml_cuda_op_gelu(ggml_backend_cuda_context & ctx, ggml_tensor * dst);
|
|
|
|
|
|
|
|
|
|
void ggml_cuda_op_silu(ggml_backend_cuda_context & ctx, ggml_tensor * dst);
|
|
|
|
|
|
2025-01-16 16:43:38 +01:00
|
|
|
void ggml_cuda_op_silu_back(ggml_backend_cuda_context & ctx, ggml_tensor * dst);
|
2025-05-24 13:06:47 +02:00
|
|
|
|
|
|
|
|
void ggml_cuda_op_gelu_erf(ggml_backend_cuda_context & ctx, ggml_tensor * dst);
|
2025-01-16 16:43:38 +01:00
|
|
|
|
2024-03-25 13:50:23 +01:00
|
|
|
void ggml_cuda_op_gelu_quick(ggml_backend_cuda_context & ctx, ggml_tensor * dst);
|
|
|
|
|
|
|
|
|
|
void ggml_cuda_op_tanh(ggml_backend_cuda_context & ctx, ggml_tensor * dst);
|
|
|
|
|
|
|
|
|
|
void ggml_cuda_op_relu(ggml_backend_cuda_context & ctx, ggml_tensor * dst);
|
|
|
|
|
|
2024-05-01 14:44:26 -07:00
|
|
|
void ggml_cuda_op_sigmoid(ggml_backend_cuda_context & ctx, ggml_tensor * dst);
|
|
|
|
|
|
2024-03-25 13:50:23 +01:00
|
|
|
void ggml_cuda_op_hardsigmoid(ggml_backend_cuda_context & ctx, ggml_tensor * dst);
|
|
|
|
|
|
2024-09-22 10:29:12 +08:00
|
|
|
void ggml_cuda_op_exp(ggml_backend_cuda_context & ctx, ggml_tensor * dst);
|
|
|
|
|
|
2024-03-25 13:50:23 +01:00
|
|
|
void ggml_cuda_op_hardswish(ggml_backend_cuda_context & ctx, ggml_tensor * dst);
|
|
|
|
|
|
|
|
|
|
void ggml_cuda_op_leaky_relu(ggml_backend_cuda_context & ctx, ggml_tensor * dst);
|
|
|
|
|
|
|
|
|
|
void ggml_cuda_op_sqr(ggml_backend_cuda_context & ctx, ggml_tensor * dst);
|
2024-06-16 15:23:04 -07:00
|
|
|
|
|
|
|
|
void ggml_cuda_op_sqrt(ggml_backend_cuda_context & ctx, ggml_tensor * dst);
|
2024-08-27 22:01:45 +03:00
|
|
|
|
|
|
|
|
void ggml_cuda_op_sin(ggml_backend_cuda_context & ctx, ggml_tensor * dst);
|
|
|
|
|
|
|
|
|
|
void ggml_cuda_op_cos(ggml_backend_cuda_context & ctx, ggml_tensor * dst);
|
2025-02-28 12:34:39 +05:30
|
|
|
|
|
|
|
|
void ggml_cuda_op_log(ggml_backend_cuda_context & ctx, ggml_tensor * dst);
|
2025-06-29 11:04:10 +02:00
|
|
|
|
2025-07-13 02:33:16 -07:00
|
|
|
void ggml_cuda_op_elu(ggml_backend_cuda_context & ctx, ggml_tensor * dst);
|
|
|
|
|
|
2025-06-29 11:04:10 +02:00
|
|
|
void ggml_cuda_op_reglu(ggml_backend_cuda_context & ctx, ggml_tensor * dst);
|
|
|
|
|
|
|
|
|
|
void ggml_cuda_op_geglu(ggml_backend_cuda_context & ctx, ggml_tensor * dst);
|
|
|
|
|
|
|
|
|
|
void ggml_cuda_op_swiglu(ggml_backend_cuda_context & ctx, ggml_tensor * dst);
|
2025-07-03 23:07:22 +02:00
|
|
|
|
|
|
|
|
void ggml_cuda_op_geglu_erf(ggml_backend_cuda_context & ctx, ggml_tensor * dst);
|
|
|
|
|
|
|
|
|
|
void ggml_cuda_op_geglu_quick(ggml_backend_cuda_context & ctx, ggml_tensor * dst);
|