#[allow(unused_imports)] use std::cmp::*; #[allow(unused_imports)] use std::collections::*; use std::io::Read; fn get_word() -> String { let stdin = std::io::stdin(); let mut stdin=stdin.lock(); let mut u8b: [u8; 1] = [0]; loop { let mut buf: Vec = Vec::with_capacity(16); loop { let res = stdin.read(&mut u8b); if res.unwrap_or(0) == 0 || u8b[0] <= b' ' { break; } else { buf.push(u8b[0]); } } if buf.len() >= 1 { let ret = String::from_utf8(buf).unwrap(); return ret; } } } fn get() -> T { get_word().parse().ok().unwrap() } /// Verified by https://atcoder.jp/contests/abc198/submissions/21774342 mod mod_int { use std::ops::*; pub trait Mod: Copy { fn m() -> i64; } #[derive(Copy, Clone, Hash, PartialEq, Eq, PartialOrd, Ord)] pub struct ModInt { pub x: i64, phantom: ::std::marker::PhantomData } impl ModInt { // x >= 0 pub fn new(x: i64) -> Self { ModInt::new_internal(x % M::m()) } fn new_internal(x: i64) -> Self { ModInt { x: x, phantom: ::std::marker::PhantomData } } pub fn pow(self, mut e: i64) -> Self { debug_assert!(e >= 0); let mut sum = ModInt::new_internal(1); let mut cur = self; while e > 0 { if e % 2 != 0 { sum *= cur; } cur *= cur; e /= 2; } sum } #[allow(dead_code)] pub fn inv(self) -> Self { self.pow(M::m() - 2) } } impl Default for ModInt { fn default() -> Self { Self::new_internal(0) } } impl>> Add for ModInt { type Output = Self; fn add(self, other: T) -> Self { let other = other.into(); let mut sum = self.x + other.x; if sum >= M::m() { sum -= M::m(); } ModInt::new_internal(sum) } } impl>> Sub for ModInt { type Output = Self; fn sub(self, other: T) -> Self { let other = other.into(); let mut sum = self.x - other.x; if sum < 0 { sum += M::m(); } ModInt::new_internal(sum) } } impl>> Mul for ModInt { type Output = Self; fn mul(self, other: T) -> Self { ModInt::new(self.x * other.into().x % M::m()) } } impl>> AddAssign for ModInt { fn add_assign(&mut self, other: T) { *self = *self + other; } } impl>> SubAssign for ModInt { fn sub_assign(&mut self, other: T) { *self = *self - other; } } impl>> MulAssign for ModInt { fn mul_assign(&mut self, other: T) { *self = *self * other; } } impl Neg for ModInt { type Output = Self; fn neg(self) -> Self { ModInt::new(0) - self } } impl ::std::fmt::Display for ModInt { fn fmt(&self, f: &mut ::std::fmt::Formatter) -> ::std::fmt::Result { self.x.fmt(f) } } impl ::std::fmt::Debug for ModInt { fn fmt(&self, f: &mut ::std::fmt::Formatter) -> ::std::fmt::Result { let (mut a, mut b, _) = red(self.x, M::m()); if b < 0 { a = -a; b = -b; } write!(f, "{}/{}", a, b) } } impl From for ModInt { fn from(x: i64) -> Self { Self::new(x) } } // Finds the simplest fraction x/y congruent to r mod p. // The return value (x, y, z) satisfies x = y * r + z * p. fn red(r: i64, p: i64) -> (i64, i64, i64) { if r.abs() <= 10000 { return (r, 1, 0); } let mut nxt_r = p % r; let mut q = p / r; if 2 * nxt_r >= r { nxt_r -= r; q += 1; } if 2 * nxt_r <= -r { nxt_r += r; q -= 1; } let (x, z, y) = red(nxt_r, r); (x, y - q * z, z) } } // mod mod_int macro_rules! define_mod { ($struct_name: ident, $modulo: expr) => { #[derive(Copy, Clone, PartialEq, Eq, PartialOrd, Ord, Hash)] pub struct $struct_name {} impl mod_int::Mod for $struct_name { fn m() -> i64 { $modulo } } } } const MOD: i64 = 998_244_353; define_mod!(P, MOD); type MInt = mod_int::ModInt

; fn convolution(a: &[MInt], b: &[MInt]) -> Vec { if a.is_empty() || b.is_empty() { return vec![]; } let n = a.len() - 1; let m = b.len() - 1; let mut ans = vec![MInt::new(0); n + m + 1]; for i in 0..n + 1 { for j in 0..m + 1 { ans[i + j] += a[i] * b[j]; } } ans } // Finds [x^n] p(x)/q(x) // Ref: https://qiita.com/ryuhe1/items/da5acbcce4ac1911f47a // Verified by: https://atcoder.jp/contests/tdpc/submissions/24583334 // Depends on: MInt.rs fn bostan_mori(p: &[MInt], q: &[MInt], mut n: i64) -> MInt { if p.is_empty() { return 0.into(); } assert!(p.len() < q.len()); let mut p = p.to_vec(); let mut q = q.to_vec(); while n > 0 { let mut qn = q.clone(); for i in 0..qn.len() { if i % 2 == 1 { qn[i] = -qn[i]; } } let num = convolution(&p, &qn); let den = convolution(&q, &qn); let mut nxt_p = vec![MInt::new(0); q.len() - 1]; let mut nxt_q = vec![MInt::new(0); q.len()]; for i in 0..q.len() - 1 { let to = 2 * i + (n % 2) as usize; if to < num.len() { nxt_p[i] = num[to]; } } for i in 0..q.len() { nxt_q[i] = den[2 * i]; } p = nxt_p; q = nxt_q; n /= 2; } p[0] * q[0].inv() } // Verified by: yukicoder No.1112 // https://yukicoder.me/submissions/510746 // https://en.wikipedia.org/wiki/Berlekamp%E2%80%93Massey_algorithm // Complexity: O(n^2) // Depends on MInt.rs fn berlekamp_massey( n: usize, s: &[mod_int::ModInt

], ) -> Vec>{ type ModInt

= mod_int::ModInt

; let mut b = ModInt::new(1); let mut cp = vec![ModInt::new(0); n + 1]; let mut bp = vec![mod_int::ModInt::new(0); n]; cp[0] = mod_int::ModInt::new(1); bp[0] = mod_int::ModInt::new(1); let mut m = 1; let mut l = 0; for i in 0..2 * n + 1 { assert!(i >= l); assert!(l <= n); if i == 2 * n { break; } let mut d = s[i]; for j in 1..l + 1 { d += cp[j] * s[i - j]; } if d == ModInt::new(0) { m += 1; continue; } if 2 * l > i { // cp -= d/b * x^m * bp let factor = d * b.inv(); for j in 0..n + 1 - m { cp[m + j] -= factor * bp[j]; } m += 1; continue; } let factor = d * b.inv(); let tp = cp.clone(); for j in 0..n + 1 - m { cp[m + j] -= factor * bp[j]; } bp = tp; b = d; l = i + 1 - l; m = 1; } cp[0..l + 1].to_vec() } // Finds u a^e v^T by using Berlekamp-massey algorithm. // The linear map a is given as a closure. // Complexity: O(n^2 log e + nT(n)) where n = |u| and T(n) = complexity of a. // Ref: https://yukicoder.me/wiki/black_box_linear_algebra // Verified by: yukicoder No. 215 (https://yukicoder.me/submissions/854179) // Depends on: fps/bostan_mori.rs fn eval_matpow Vec>(mut a: F, e: i64, u: &[MInt], v: &[MInt]) -> MInt { let k = u.len(); // Find first 2k terms let mut terms = vec![MInt::new(0); 2 * k]; let mut cur = u.to_vec(); for pos in 0..2 * k { for i in 0..k { terms[pos] += cur[i] * v[i]; } cur = a(&cur); } let poly = berlekamp_massey(k, &terms); let mut nom = convolution(&terms[..poly.len() - 1], &poly); nom.truncate(poly.len() - 1); bostan_mori(&nom, &poly, e) } // #{x | x < m, (x as bitstring)[p] = 1} fn count_pop_bits(m: i64, p: usize) -> i64 { let lead = m & ((-1) << (p + 1)); let rest = m - lead; let ans = (lead >> 1) + if rest >= 1 << p { rest - (1 << p) } else { 0 }; ans } // \sum_{x < m} (x xor 2^p) - x pub fn c(m: i64, p: usize) -> MInt { let tmp = MInt::new(m) - count_pop_bits(m, p) * 2; tmp * MInt::new(2).pow(p as i64) } // \sum{x | x < m, (x as bitstring)[p] = 1} pub fn e(m: i64, p: usize) -> MInt { let lead = m & ((-1) << (p + 1)); let rest = m - lead; let p2 = MInt::new(1 << p); let inv2 = MInt::new(2).inv(); let count = MInt::new(lead >> (p + 1)); let mut tot = p2 * (p2 * 3 - 1) * inv2 * count; tot += count * (count - 1) * inv2 * MInt::new(1 << (p + 1)) * (1 << p); if rest >= 1 << p { let tmp = MInt::new(rest - (1 << p)); tot += MInt::new(lead + (1 << p)) * tmp; tot += tmp * (tmp - 1) * inv2; } tot } // \sum_{k < m, (k as bitstring)[p] = 1} (2^i xor k) - k pub fn d(m: i64, p: usize, i: usize) -> MInt { if p == i { return -MInt::new(count_pop_bits(m, p)) * (1 << i); } let lead = m & ((-1) << (p + 1)); let rest = m - lead; if i < p { return if rest >= 1 << p { c(rest - (1 << p), i) } else { 0.into() }; } let mut ans = c(lead >> (p + 1), i - p - 1) * (1 << p) * (1 << (p + 1)); let tmp = if rest >= 1 << p { MInt::new(rest - (1 << p)) * (1 << i) } else { MInt::new(0) }; if (lead & (1 << i)) != 0 { ans -= tmp; } else { ans += tmp; } ans } // dp[i][j] := i 番目まで埋めて A_i= j のときの積の総和 とすると、dp[i] |-> dp[i+1] は線型変換。 // これを行列累乗する必要があるが、そのままだと次元が M であり大きすぎるのである程度まとめる必要がある。 // u_j = dp[i][j], v_j = dp[i+1][j] として、u から v を作る線型変換のより小さい不変部分空間を作る。 // v_j = \sum_k u_k (k xor j) である。 // a := (u_j の和), s_i := {i 番目ビットが立っているもの限定の u_j の和}, // b := (v_j の和), t_i := {i 番目ビットが立っているもの限定の v_j の和} とする。 // a, s_i から b, t_i が計算できるのがポイント。そのためには以下の補題を使う。 // 補題: k = 2^a + 2^b + ... とする。このとき (k xor j) - j = ((2^a xor j) - j) + ((2^b xor j) - j) + ... // この補題を使うと、まず b = (M(M-1)/2) a + \sum c(2^i) s_i が言える。(c(x) := \sum_{j