#include <cuda_fp16.h>
#include <cuda_bf16.h>
#include <hidet/runtime/cuda_context.h>
#include <hidet/runtime/cpu_context.h>
typedef float tfloat32_t;
#define __float_to_tf32(x) (x)
/*
Task(
  name: clip
  parameters:
    x: tensor(int64, [1, 4, 1, 1])
    y: tensor(int64, [1, 4, 1, 1])
  inputs: [x]
  outputs: [y]
  computations:
    x: tensor(int64, [1, 4, 1, 1])
    y: int64[1, 4, 1, 1] where y[v, v_1, v_2, v_3] = generic_min(generic_max(x[v, v_1, v_2, v_3], int64(-1)), int64(1))
  attributes: {}
)
*/
extern "C" {

__global__ void __launch_bounds__(500) hidet_compute_y(int64_t * __restrict__ x, int64_t * __restrict__ y) {
  if ((int)threadIdx.x < 4) {
    y[((int)threadIdx.x % 4)] = min(max(x[((int)threadIdx.x % 4)], -1ll), 1ll);
  }
}

__host__ void hidet_clip(int32_t num_args, int32_t * __restrict__ arg_types, void* * __restrict__ args) {
  hidet_compute_y<<<1, 500, 0, (cudaStream_t)get_cuda_stream()>>>(((int64_t*)(args[0])), ((int64_t*)(args[1])));
}

}

 error: more than one instance of overloaded function "max" matches the argument list:
            function "max(long, long)"
/usr/local/cuda/bin/../targets/x86_64-linux/include/crt/math_functions.hpp(1008): here
            function "max(long, unsigned long)"
/usr/local/cuda/bin/../targets/x86_64-linux/include/crt/math_functions.hpp(1043): here
            function "max(long long, long long)"
/usr/local/cuda/bin/../targets/x86_64-linux/include/crt/math_functions.hpp(1077): here
            function "max(unsigned long long, long long)"
/usr/local/cuda/bin/../targets/x86_64-linux/include/crt/math_functions.hpp(1092): here
            argument types are: (int64_t, long long)