/tmp/solutions/build/range_affine_range_sum-main.cpp:
1| |#include <common.h>
2| |#include <toy/bit.h>
3| |prelude;
4| |
5| |namespace {
6| |
7| |constexpr int N = 1e6;
8| |constexpr int P = 998244353;
9| |
10| |struct affine {
11| | u32 a, b;
12| 18.6M| auto operator+(affine t) -> affine {
13| 18.6M| return {u32(u64(t.a) * a % P), u32((u64(t.a) * b + t.b) % P)};
14| 18.6M| }
15| 18.6M| auto operator+=(affine t) -> void { *this = *this + t; }
16| |};
17| |
18| |struct node {
19| | u32 siz;
20| | u32 sum;
21| | affine aff;
22| 18.6M| auto operator+=(affine t) -> void {
23| 18.6M| aff += t;
24| 18.6M| sum = (u64(t.a) * sum + u64(t.b) * siz) % P;
25| 18.6M| }
26| |} a[N];
27| |
28| 409k|void pushdown(int k) {
29| 8.05M| for (int i = log(k) - 1; i >= 0; --i) {
^7.64M
------------------
| Branch (29:28): [True: 94.91%, False: 5.09%]
------------------
30| 7.64M| affine t = a[k >> i >> 1].aff;
31| 7.64M| a[k >> i >> 1].aff = affine{1, 0};
32| 7.64M| a[k >> i ^ 0] += t;
33| 7.64M| a[k >> i ^ 1] += t;
34| 7.64M| }
35| 409k|}
36| |
37| 7.83M|u32 mod(u32 x) { return x < P ? x : x - P; }
^3.91M^3.91M
------------------
| Branch (37:25): [True: 50.01%, False: 49.99%]
------------------
38| |
39| 409k|void pushup(int k) {
40| 7.64M| for (k /= 2; k > 0; k /= 2) {
^7.23M
------------------
| Branch (40:16): [True: 94.64%, False: 5.36%]
------------------
41| 7.23M| a[k].sum = mod(a[k * 2].sum + a[k * 2 + 1].sum);
42| 7.23M| }
43| 409k|}
44| |
45| |}; // namespace
46| |
47| 1|int main() {
48| 1| rd rd;
49| 1| wt wt;
50| 1| int n = rd.uh();
51| 1| int q = rd.uh();
52| 389k| for (int i = 0; i < n; ++i) a[n + i].siz = 1, a[n + i].sum = rd.uw();
^389k^389k
------------------
| Branch (52:19): [True: 100.00%, False: 0.00%]
------------------
53| 389k| for (int i = n - 1; i >= 0; --i) {
^389k
------------------
| Branch (53:23): [True: 100.00%, False: 0.00%]
------------------
54| 389k| a[i].aff = {1, 0};
55| 389k| a[i].siz = a[2 * i].siz + a[2 * i + 1].siz;
56| 389k| a[i].sum = mod(a[2 * i].sum + a[2 * i + 1].sum);
57| 389k| }
58| 410k| while (q--) {
------------------
| Branch (58:10): [True: 100.00%, False: 0.00%]
------------------
59| 410k| let t = rd.u1();
60| 410k| if (t == 0) {
------------------
| Branch (60:9): [True: 49.88%, False: 50.12%]
------------------
61| 204k| int l = n + rd.uh();
62| 204k| int r = n + rd.uh() - 1;
63| 204k| pushdown(l--);
64| 204k| pushdown(r++);
65| 204k| int k = log(l ^ r);
66| 204k| affine x = {rd.uw(), rd.uw()};
67| 204k| int t, i;
68| 1.87M| for (t = ~l & ~(-1 << k), i = 31; t > 0; t -= 1 << i) {
^1.66M
------------------
| Branch (68:41): [True: 89.06%, False: 10.94%]
------------------
69| 1.66M| i = log(t);
70| 1.66M| a[l >> i ^ 1] += x;
71| 1.66M| }
72| 204k| pushup(l >> i);
73| 1.91M| for (t = +r & ~(-1 << k), i = 31; t > 0; t -= 1 << i) {
^1.70M
------------------
| Branch (73:41): [True: 89.27%, False: 10.73%]
------------------
74| 1.70M| i = log(t);
75| 1.70M| a[r >> i ^ 1] += x;
76| 1.70M| }
77| 204k| pushup(r >> i);
78| 204k| }
79| 410k| if (t == 1) {
------------------
| Branch (79:9): [True: 50.12%, False: 49.88%]
------------------
80| 205k| int l = n + rd.uh();
81| 205k| int r = n + rd.uh() - 1;
82| 205k| u64 sizL = 0, sumL = 0;
83| 205k| u64 sizR = 0, sumR = 0;
84| 3.75M| for (--l, ++r; l ^ r ^ 1;) {
------------------
| Branch (84:22): [True: 94.52%, False: 5.48%]
------------------
85| 3.55M| if (~l & 1) sizL += a[l ^ 1].siz, sumL += a[l ^ 1].sum;
^1.67M
------------------
| Branch (85:13): [True: 47.14%, False: 52.86%]
------------------
86| 3.55M| if (+r & 1) sizR += a[r ^ 1].siz, sumR += a[r ^ 1].sum;
^1.71M
------------------
| Branch (86:13): [True: 48.26%, False: 51.74%]
------------------
87| 3.55M| l /= 2, r /= 2;
88| 3.55M| sumL = (a[l].aff.a * sumL + a[l].aff.b * sizL) % P;
89| 3.55M| sumR = (a[r].aff.a * sumR + a[r].aff.b * sizR) % P;
90| 3.55M| }
91| 205k| sumL = mod(u32(sumL + sumR));
92| 205k| sizL += sizR;
93| 543k| for (l /= 2; l > 0; l /= 2) {
^337k
------------------
| Branch (93:20): [True: 62.12%, False: 37.88%]
------------------
94| 337k| sumL = (a[l].aff.a * sumL + a[l].aff.b * sizL) % P;
95| 337k| }
96| 205k| wt.uw(u32(sumL));
97| 205k| }
98| 410k| }
99| 1| return 0;
100| 1|}