#include <bits/stdc++.h>

int ri() {
	int n;
	scanf("%d", &n);
	return n;
}

template<int mod>
struct ModMatrixPower {
	std::vector<std::vector<int> > data;
	std::vector<std::vector<int> > tmp;
	ModMatrixPower(int n) {
		data.resize(n, std::vector<int>(n));
		tmp.resize(n, std::vector<int>(n));
	}
	auto & operator [] (int x) { return data[x]; }
	void pow_self() {
		size_t n = data.size();
		int r0[n][n], r1[n][n];
		for (size_t i = 0; i < n; i++) for (size_t j = 0; j < n; j++) {
			r0[i][j] = data[i][j];
			r1[j][i] = data[i][j];
		}
		constexpr size_t BLOCK = 0x8;
		for (size_t i = 0; i < n; i += BLOCK) {
			for (size_t j = 0; j < n; j += BLOCK) {
				for (size_t k = i; k < std::min(i + BLOCK, n); k++) {
					for (size_t l = j; l < std::min(j + BLOCK, n); l++) {
						uint64_t sum = 0;
						size_t upper1 = n & 0x7FFFFFF0;
						for (size_t m = 0; m < upper1; m += 0x10) {
							sum += (uint64_t) r0[k][m] * r1[l][m];
							sum += (uint64_t) r0[k][m + 1] * r1[l][m + 1];
							sum += (uint64_t) r0[k][m + 2] * r1[l][m + 2];
							sum += (uint64_t) r0[k][m + 3] * r1[l][m + 3];
							sum += (uint64_t) r0[k][m + 4] * r1[l][m + 4];
							sum += (uint64_t) r0[k][m + 5] * r1[l][m + 5];
							sum += (uint64_t) r0[k][m + 6] * r1[l][m + 6];
							sum += (uint64_t) r0[k][m + 7] * r1[l][m + 7];
							sum += (uint64_t) r0[k][m + 8] * r1[l][m + 8];
							sum += (uint64_t) r0[k][m + 9] * r1[l][m + 9];
							sum += (uint64_t) r0[k][m + 10] * r1[l][m + 10];
							sum += (uint64_t) r0[k][m + 11] * r1[l][m + 11];
							sum += (uint64_t) r0[k][m + 12] * r1[l][m + 12];
							sum += (uint64_t) r0[k][m + 13] * r1[l][m + 13];
							sum += (uint64_t) r0[k][m + 14] * r1[l][m + 14];
							sum += (uint64_t) r0[k][m + 15] * r1[l][m + 15];
							sum %= mod;
						}
						for (size_t m = upper1; m < n; m++) sum += (uint64_t) r0[k][m] * r1[l][m];
						tmp[k][l] = sum % mod;
					}
				}
			}
		}
		std::swap(tmp, data);
	}
	void apply(std::vector<int> &vec) {
		size_t n = data.size();
		assert(vec.size() == n);
		std::vector<int> vec_tmp(n);
		for (size_t i = 0; i < n; i++)
			for (size_t j = 0; j < n; j++) {
				vec_tmp[j] += (int64_t) vec[i] * data[i][j] % mod;
				if (vec_tmp[j] >= mod) vec_tmp[j] -= mod;
			}
		std::swap(vec, vec_tmp);
	}
};

#define MOD 998244353

int main() {
	int n = ri();
	int k = ri();
	int kk = k * k;
	int kkk = k * k * k;
	ModMatrixPower<MOD> mat(kkk);
	for (int i = 0; i < kkk; i++) {
		int x = i / kk;
		int y = i / k % k;
		int z = i % k;
		mat[i][x * kk + y * k + (z + 1) % k]++;
		mat[i][x * kk + (y + z) % k * k + z]++;
		mat[i][(x + y) % k * kk + y * k + z]++;
	}
	std::vector<int> vec(kkk, 0);
	vec[0] = 1;
	
	for (; n; n >>= 1) {
		if (n & 1) mat.apply(vec);
		mat.pow_self();
	}
	int res = 0;
	for (int i = 0; i < kk; i++) {
		res += vec[i];
		if (res >= MOD) res -= MOD;
	}
	printf("%d\n", res);
	
	return 0;
}