#ifdef CK_EXPERIMENTAL_BIT_INT_EXTENSION_INT4

#include "common.hpp"

#include "ck/tensor_operation/gpu/device/impl/device_gemm_dl.hpp"

using ADataType = ck::int4_t;

using BDataType = ck::int4_t;

using CDataType = ck::int4_t;

using KernelADataType = int8_t;

using KernelBDataType = int8_t;

using KernelCDataType = int8_t;

using AccDataType = int32_t;

using ALayout = Col;

using BLayout = Row;

using CLayout = Row;

using AElementOp = PassThrough;

using BElementOp = PassThrough;

using CElementOp = PassThrough;

static constexpr auto GemmDefault = ck::tensor_operation::device::GemmSpecialization::Default;

using DeviceGemmInstance = ck::tensor_operation::device::DeviceGemmDl

< KernelADataType, KernelBDataType, KernelCDataType, AccDataType, ALayout, BLayout, CLayout, AElementOp, BElementOp, CElementOp, GemmDefault, 256, 128, 128, 16, 4, 4, 4, 1, S<8, 2>, S<8, 2>, S<2, 1, 4, 4>, S<8, 1, 32, 1>, S<0, 3, 1, 2>, S<0, 3, 1, 2>, S<1, 1, 4, 1>, S<0, 3, 1, 2>, S<1, 1, 4, 4>, S<2, 1, 4, 4>, S<8, 1, 32, 1>, S<0, 3, 1, 2>, S<0, 3, 1, 2>, S<1, 1, 4, 1>, S<0, 3, 1, 2>, S<1, 1, 4, 4>, S<0, 1, 2, 3, 4, 5>, 5, 4>;

using ReferenceGemmInstance = ck::tensor_operation::host::

ReferenceGemm<ADataType, BDataType, CDataType, AccDataType, AElementOp, BElementOp, CElementOp>;

#define BUILD_INT4_EXAMPLE

#include "run_gemm_example.inc"

int main(int argc, char* argv[]) { return !run_gemm_example(argc, argv); }

#endif