import torch
import torch.nn as nn

# Default ModelNew is PyTorch-only. run_and_check: use check_kernel=False with backend=cuda
# until the candidate file contains a real CUDA kernel (see kernel_static_checker).

class Model(nn.Module):
    """
    Fused 1D convolution + GroupNorm for temporal sequences (N, C, T).

    Typical in time-series heads, audio / motor trajectories, and sequence models
    that use strided or padded Conv1d followed by per-channel normalization without
    collapsing the time dimension (GroupNorm over `out_channels` on (N, C_out, T')).

    Reference fusion target: `Conv1d` → `GroupNorm` (no activation), matching a
    common stem or residual branch in temporal backbones.
    """

    def __init__(
        self,
        in_channels: int,
        out_channels: int,
        num_groups: int,
        kernel_size: int = 3,
        stride: int = 1,
        padding: int = 1,
        dilation: int = 1,
        eps: float = 1e-5,
        affine: bool = True,
    ):
        super().__init__()
        assert out_channels % num_groups == 0
        self.conv = nn.Conv1d(
            in_channels,
            out_channels,
            kernel_size,
            stride=stride,
            padding=padding,
            dilation=dilation,
            bias=False,
        )
        self.norm = nn.GroupNorm(
            num_groups=num_groups,
            num_channels=out_channels,
            eps=eps,
            affine=affine,
        )

    def forward(self, x: torch.Tensor) -> torch.Tensor:
        """
        Args:
            x: (N, C_in, T) temporal feature map
        Returns:
            (N, C_out, T') with T' determined by conv stride / padding / kernel
        """
        return self.norm(self.conv(x))


class ModelNew(nn.Module):
    """
    KernelBench candidate entry point. Replace with a fused CUDA/Triton kernel;
    default delegates to `Model` for harness smoke tests.
    """

    def __init__(
        self,
        in_channels: int,
        out_channels: int,
        num_groups: int,
        kernel_size: int = 3,
        stride: int = 1,
        padding: int = 1,
        dilation: int = 1,
        eps: float = 1e-5,
        affine: bool = True,
    ):
        super().__init__()
        self._impl = Model(
            in_channels,
            out_channels,
            num_groups,
            kernel_size=kernel_size,
            stride=stride,
            padding=padding,
            dilation=dilation,
            eps=eps,
            affine=affine,
        )

    def forward(self, x: torch.Tensor) -> torch.Tensor:
        return self._impl(x)


batch_size = 16
in_channels = 128
out_channels = 256
num_groups = 32
seq_len = 1024
kernel_size = 3
stride = 1
padding = 1


def get_inputs():
    p = popcorn_pri
    x = torch.randn(p.jitter_int(batch_size), in_channels, p.jitter_int(seq_len))
    return [x]


def get_init_inputs():
    return [
        in_channels,
        out_channels,
        num_groups,
        kernel_size,
        stride,
        padding,
    ]

import torch
import torch.nn as nn
import torch.nn.functional as F
from torch.utils.cpp_extension import load_inline

cuda_source = r'''
#include <torch/extension.h>
#include <cuda_runtime.h>

__global__ void groupnorm_kernel(
    const float* __restrict__ input,
    const float* __restrict__ gamma,
    const float* __restrict__ beta,
    float* __restrict__ output,
    int N, int C, int T, int num_groups, float eps
) {
    int G = num_groups;
    int C_per_g = C / G;
    int ng = blockIdx.x;
    int n = ng / G;
    int g = ng % G;
    
    float sum = 0.0f;
    float sq_sum = 0.0f;
    for (int c = 0; c < C_per_g; ++c) {
        int ch = g * C_per_g + c;
        for (int t = 0; t < T; ++t) {
            float val = input[((n * C + ch) * T) + t];
            sum += val;
            sq_sum += val * val;
        }
    }
    float mean = sum / (C_per_g * T);
    float var = sq_sum / (C_per_g * T) - mean * mean;
    float inv_std = rsqrtf(var + eps);
    
    for (int c = 0; c < C_per_g; ++c) {
        int ch = g * C_per_g + c;
        float g_ = gamma[ch];
        float b = beta[ch];
        for (int t = 0; t < T; ++t) {
            float val = input[((n * C + ch) * T) + t];
            float normed = (val - mean) * inv_std * g_ + b;
            output[((n * C + ch) * T) + t] = normed;
        }
    }
}

torch::Tensor groupnorm_cuda(
    torch::Tensor input,
    torch::Tensor gamma,
    torch::Tensor beta,
    int num_groups, float eps
) {
    int N = input.size(0);
    int C = input.size(1);
    int T = input.size(2);
    auto output = torch::empty_like(input);
    
    int G = num_groups;
    int blocks = N * G;
    
    groupnorm_kernel<<<blocks, 1>>>(
        input.data_ptr<float>(),
        gamma.data_ptr<float>(),
        beta.data_ptr<float>(),
        output.data_ptr<float>(),
        N, C, T, num_groups, eps
    );
    
    return output;
}
'''

cpp_source = r'''
torch::Tensor groupnorm_cuda(
    torch::Tensor input,
    torch::Tensor gamma,
    torch::Tensor beta,
    int num_groups, float eps
);
'''

module = load_inline(
    name='groupnorm_ext',
    cpp_sources=[cpp_source],
    cuda_sources=[cuda_source],
    functions=['groupnorm_cuda'],
    verbose=True,
    extra_cuda_cflags=['-O3']
)

class ModelNew(nn.Module):
    def __init__(self, in_channels, out_channels, num_groups, kernel_size=3, stride=1, padding=1, dilation=1, eps=1e-5, affine=True):
        super().__init__()
        self.conv = nn.Conv1d(in_channels, out_channels, kernel_size, stride=stride, padding=padding, dilation=dilation, bias=False)
        self.norm = nn.GroupNorm(num_groups, out_channels, eps=eps, affine=affine)
        self.num_groups = num_groups
        self.eps = eps

    def forward(self, x):
        out = F.conv1d(x, self.conv.weight, None, self.conv.stride, self.conv.padding, self.conv.dilation)
        return module.groupnorm_cuda(out, self.norm.weight, self.norm.bias, self.num_groups, self.eps)