#include <torch/library.h>

#include "registration.h"
#include "torch_binding.h"

TORCH_LIBRARY_EXPAND(TORCH_EXTENSION_NAME, ops) {
  // Increment a tensor by 1.
  ops.def("increment(Tensor x) -> ()");
  ops.impl("increment", torch::kCUDA, &increment);

  // Compute the residual root mean square.
  ops.def("residual_rms(Tensor input, Tensor residual, Tensor weight, Tensor output, float epsilon, float scale, int mode, int num_threads) -> ()");
  ops.impl("residual_rms", torch::kCUDA, &residual_rms);
}

REGISTER_EXTENSION(TORCH_EXTENSION_NAME)