| void ggml_cuda_op_argsort(ggml_backend_cuda_context & ctx, ggml_tensor * dst); | |
| void argsort_f32_i32_cuda_cub(ggml_cuda_pool & pool, | |
| const float * x, | |
| int * dst, | |
| const int ncols, | |
| const int nrows, | |
| ggml_sort_order order, | |
| cudaStream_t stream); | |
| void argsort_f32_i32_cuda_bitonic(const float * x, | |
| int * dst, | |
| const int ncols, | |
| const int nrows, | |
| ggml_sort_order order, | |
| cudaStream_t stream); | |