import triton
import triton.language as tl

@triton.jit
def matmul_kernel(
    a_ptr, b_ptr, c_ptr,
    M, N, K,
    stride_am, stride_ak,
    stride_bk, stride_bn,
    stride_cm, stride_cn,
    BLOCK_SIZE_M: tl.constexpr, BLOCK_SIZE_N: tl.constexpr, BLOCK_SIZE_K: tl.constexpr,
    SPLIT_K: tl.constexpr, GROUP_SIZE_M: tl.constexpr,
):
    pid = tl.program_id(axis=0)
    pid_z = tl.program_id(1)
    num_pid_m = tl.cdiv(M, BLOCK_SIZE_M)
    num_pid_n = tl.cdiv(N, BLOCK_SIZE_N)
    if GROUP_SIZE_M == 1:
        pid_m = pid // num_pid_n
        pid_n = pid % num_pid_n
    else:
        num_pid_in_group = GROUP_SIZE_M * num_pid_n
        group_id = pid // num_pid_in_group
        first_pid_m = group_id * GROUP_SIZE_M
        group_size_m = min(num_pid_m - first_pid_m, GROUP_SIZE_M)
        pid_m = first_pid_m + (pid % group_size_m)
        pid_n = (pid % num_pid_in_group) // group_size_m
    if SPLIT_K == 1:
        offs_k = tl.arange(0, BLOCK_SIZE_K)
    else:
        offs_k = pid_z * BLOCK_SIZE_K + tl.arange(0, BLOCK_SIZE_K)
    offs_am = (pid_m * BLOCK_SIZE_M + tl.arange(0, BLOCK_SIZE_M))
    offs_bn = (pid_n * BLOCK_SIZE_N + tl.arange(0, BLOCK_SIZE_N))
    a_ptrs = a_ptr + offs_am[:, None] * stride_am + offs_k[None, :] * stride_ak
    b_ptrs = b_ptr + offs_k[:, None] * stride_bk + offs_bn[None, :] * stride_bn
    acc_dtype = tl.float32 if c_ptr.type.element_ty != tl.int8 else tl.int32
    accumulator = tl.zeros((BLOCK_SIZE_M, BLOCK_SIZE_N), dtype=acc_dtype)
    for k in range(0, tl.cdiv(K, BLOCK_SIZE_K * SPLIT_K)):
        a = tl.load(a_ptrs)
        b = tl.load(b_ptrs)
        accumulator += tl.dot(a, b)
        a_ptrs += BLOCK_SIZE_K * SPLIT_K * stride_ak
        b_ptrs += BLOCK_SIZE_K * SPLIT_K * stride_bk
    c = accumulator.to(c_ptr.type.element_ty)
    offs_cm = pid_m * BLOCK_SIZE_M + tl.arange(0, BLOCK_SIZE_M)
    offs_cn = pid_n * BLOCK_SIZE_N + tl.arange(0, BLOCK_SIZE_N)
    c_ptrs = c_ptr + stride_cm * offs_cm[:, None] + stride_cn * offs_cn[None, :]
    c_mask = (offs_cm[:, None] < M) & (offs_cn[None, :] < N)
    if SPLIT_K == 1:
        tl.store(c_ptrs, c, mask=c_mask)
    else:
        tl.atomic_add(c_ptrs, c, mask=c_mask)