Problem 508
Integers in Base i-1 — B(10^15) mod 1e9+7 via recursive region sums.
View problem on Project Euler
Performance comparison
| Metric | Our solution | Best known |
|---|
| Time complexity | O(n^2) | ? |
|---|
| Space complexity | O(n^2) | ? |
|---|
| Approach | Flow solution | Not curated |
|---|
| Verdict | Unknown |
|---|
Flow source
# Project Euler 508
# Integers in Base i-1 — B(10^15) mod 1e9+7 via recursive region sums.
const MOD: i64 = 1000000007
const LIM: i64 = 1000000000000000
const BRUTE: i64 = 4000
const CAP: i64 = 1048576
extern {
function calloc(n: i64, size: i64) -> ptr<void>
function free(p: ptr<void>) -> void
}
function floordiv(a: i64, b: i64) -> i64 {
if b == 0 { return 0 }
let q: i64 = a / b
let r: i64 = a - q * b
if r != 0 && ((r > 0) != (b > 0)) { q = q - 1 }
return q
}
function f_gauss(a0: i64, b0: i64) -> i64 {
let mut a: i64 = a0
let mut b: i64 = b0
let mut cnt: i64 = 0
while a != 0 || b != 0 {
if ((a ^ b) & 1) != 0 {
a = a - 1
cnt = cnt + 1
}
let na: i64 = floordiv(b - a, 2)
let nb: i64 = floordiv(-(a + b), 2)
a = na
b = nb
}
return cnt
}
function count_rect(x0: i64, x1: i64, y0: i64, y1: i64) -> i128 {
if x0 > x1 || y0 > y1 { return 0 as i128 }
let w: i128 = ((x1 - x0 + 1) as i128)
let h: i128 = ((y1 - y0 + 1) as i128)
return w * h
}
function count_parity(lo: i64, hi: i64, parity: i64) -> i64 {
if lo > hi { return 0 }
let mut first: i64 = lo
if (lo & 1) != parity { first = lo + 1 }
if first > hi { return 0 }
return floordiv(hi - first, 2) + 1
}
function count_diamond(u0: i64, u1: i64, v0: i64, v1: i64) -> i128 {
if u0 > u1 || v0 > v1 { return 0 as i128 }
let eu: i128 = (count_parity(u0, u1, 0) as i128)
let ou: i128 = ((u1 - u0 + 1) as i128) - eu
let ev: i128 = (count_parity(v0, v1, 0) as i128)
let ov: i128 = ((v1 - v0 + 1) as i128) - ev
return eu * ev + ou * ov
}
function ceil_div(n: i64, d: i64) -> i64 {
return -floordiv(-n, d)
}
function hkey(a0: i64, a1: i64, b0: i64, b1: i64, tag: i64) -> i64 {
let P: i128 = 1000000000000000003 as i128
let mut k: i128 = (tag as i128)
k = (k * (1000003 as i128) + (a0 as i128)) % P
k = (k * (1000003 as i128) + (a1 as i128)) % P
k = (k * (1000003 as i128) + (b0 as i128)) % P
k = (k * (1000003 as i128) + (b1 as i128)) % P
if k < 0 { k = k + P }
return k as i64
}
function hslot(key: i64, keys: ptr<i64>, used: ptr<i8>) -> i64 {
let mut h: i64 = key % CAP
if h < 0 { h = h + CAP }
while used[h] == 1 && keys[h] != key {
h = h + 1
if h == CAP { h = 0 }
}
return h
}
function brute_rect(x0: i64, x1: i64, y0: i64, y1: i64) -> i64 {
let mut s: i64 = 0
let mut a: i64 = x0
while a <= x1 {
let mut b: i64 = y0
while b <= y1 {
s = s + f_gauss(a, b)
b = b + 1
}
a = a + 1
}
return s % MOD
}
function brute_diamond(u0: i64, u1: i64, v0: i64, v1: i64) -> i64 {
let mut s: i64 = 0
let mut u: i64 = u0
while u <= u1 {
let parity: i64 = u & 1
let mut v: i64 = v0
if (v & 1) != parity { v = v + 1 }
while v <= v1 {
s = s + f_gauss(floordiv(u + v, 2), floordiv(u - v, 2))
v = v + 2
}
u = u + 1
}
return s % MOD
}
function rect_to_diamond(x0: i64, x1: i64, y0: i64, y1: i64, r: i64,
u0: ptr<i64>, u1: ptr<i64>, v0: ptr<i64>, v1: ptr<i64>) -> void {
u0[0] = -x1 + r
u1[0] = -x0 + r
v0[0] = y0
v1[0] = y1
}
function diamond_to_rect(u0: i64, u1: i64, v0: i64, v1: i64, r: i64,
x0: ptr<i64>, x1: ptr<i64>, y0: ptr<i64>, y1: ptr<i64>) -> void {
x0[0] = ceil_div(r - v1, 2)
x1[0] = floordiv(r - v0, 2)
y0[0] = ceil_div(r - u1, 2)
y1[0] = floordiv(r - u0, 2)
}
function sum_region(tag: i32, a0: i64, a1: i64, b0: i64, b1: i64,
keys: ptr<i64>, vals: ptr<i64>, used: ptr<i8>, buf: ptr<i64>) -> i64 {
if a0 > a1 || b0 > b1 { return 0 }
if tag == 0 {
let n0: i128 = count_rect(a0, a1, b0, b1)
if n0 <= (BRUTE as i128) { return brute_rect(a0, a1, b0, b1) }
} else {
let n1: i128 = count_diamond(a0, a1, b0, b1)
if n1 <= (BRUTE as i128) { return brute_diamond(a0, a1, b0, b1) }
}
let key: i64 = hkey(a0, a1, b0, b1, tag as i64)
let slot: i64 = hslot(key, keys, used)
if used[slot] == 1 && keys[slot] == key { return vals[slot] }
let mut res: i64 = 0
if tag == 0 {
rect_to_diamond(a0, a1, b0, b1, 0, buf, buf + 1, buf + 2, buf + 3)
res = sum_region(1, buf[0], buf[1], buf[2], buf[3], keys, vals, used, buf)
rect_to_diamond(a0, a1, b0, b1, 1, buf, buf + 1, buf + 2, buf + 3)
let du0: i64 = buf[0]
let du1: i64 = buf[1]
let dv0: i64 = buf[2]
let dv1: i64 = buf[3]
res = (res + sum_region(1, du0, du1, dv0, dv1, keys, vals, used, buf)) % MOD
res = (res + ((count_diamond(du0, du1, dv0, dv1) % (MOD as i128)) as i64)) % MOD
} else {
diamond_to_rect(a0, a1, b0, b1, 0, buf, buf + 1, buf + 2, buf + 3)
res = sum_region(0, buf[0], buf[1], buf[2], buf[3], keys, vals, used, buf)
diamond_to_rect(a0, a1, b0, b1, 1, buf, buf + 1, buf + 2, buf + 3)
let rx0: i64 = buf[0]
let rx1: i64 = buf[1]
let ry0: i64 = buf[2]
let ry1: i64 = buf[3]
res = (res + sum_region(0, rx0, rx1, ry0, ry1, keys, vals, used, buf)) % MOD
res = (res + ((count_rect(rx0, rx1, ry0, ry1) % (MOD as i128)) as i64)) % MOD
}
let store: i64 = hslot(key, keys, used)
used[store] = 1
keys[store] = key
vals[store] = res
return res
}
function main() -> i32 {
let keys: ptr<i64> = calloc(CAP, 8)
let vals: ptr<i64> = calloc(CAP, 8)
let used: ptr<i8> = calloc(CAP, 1)
let buf: ptr<i64> = calloc(4, 8)
if keys == null || vals == null || used == null || buf == null { return 1 }
let ans: i64 = sum_region(0, -LIM, LIM, -LIM, LIM, keys, vals, used, buf)
printf("%lld\n", ans)
free(keys); free(vals); free(used); free(buf)
return 0
}
Generated C
#include <stdint.h>
#include <stdbool.h>
#include <stdio.h>
#include <stdlib.h>
#include <string.h>
/* Flow runtime helpers */
typedef struct flow_temp_node { struct flow_temp_node* next; } flow_temp_node;
static flow_temp_node* flow_temp_head = NULL;
static int flow_temp_atexit_set = 0;
__attribute__((unused)) static void flow_temp_free_all(void) {
while (flow_temp_head) {
flow_temp_node* n = flow_temp_head;
flow_temp_head = n->next;
free(n);
}
}
__attribute__((unused)) static void* flow_temp_alloc(size_t nbytes) {
flow_temp_node* node = (flow_temp_node*)malloc(sizeof(flow_temp_node) + nbytes);
if (!node) return NULL;
node->next = flow_temp_head;
flow_temp_head = node;
if (!flow_temp_atexit_set) {
flow_temp_atexit_set = 1;
atexit(flow_temp_free_all);
}
return (void*)(node + 1);
}
#ifndef FLOW_DIAG
#define FLOW_DIAG(msg) fprintf(stderr, "%s", (msg))
#endif
#ifndef FLOW_LOG
#define FLOW_LOG(fmt, ...) printf(fmt, __VA_ARGS__)
#endif
#ifndef FLOW_LOG_EMPTY
#define FLOW_LOG_EMPTY(fmt) printf(fmt)
#endif
static char* flow_strcat(const char* a, const char* b) {
size_t la = strlen(a ? a : ""), lb = strlen(b ? b : "");
char* r = (char*)flow_temp_alloc(la + lb + 1);
if (!r) return NULL;
if (la) memcpy(r, a, la);
if (lb) memcpy(r + la, b, lb);
r[la + lb] = '\0';
return r;
}
#define __flow_in_arr(arr, val) __extension__ ({ \
int _found = 0; \
size_t _n = sizeof(arr)/sizeof((arr)[0]); \
for (size_t _i = 0; _i < _n; _i++) { \
if ((arr)[_i] == (val)) { _found = 1; break; } \
} _found; })
/* Unified fault handler (MISRA #279) — override with -DFLOW_FAULT_HANDLER=fn */
#ifndef FLOW_FAULT_HANDLER
__attribute__((unused)) static inline void flow_fault_handler(const char* msg) {
fprintf(stderr, "flow: %s\n", msg ? msg : "fault");
abort();
#if defined(__GNUC__) || defined(__clang__)
__builtin_unreachable();
#endif
}
#else
#define flow_fault_handler FLOW_FAULT_HANDLER
#endif
#define flow_div_by_zero_handler() flow_fault_handler("division by zero")
#define flow_shift_ub_handler() flow_fault_handler("invalid shift (amount out of range or left-shift of negative)")
#ifndef FLOW_CHECKED_DIV
#define FLOW_CHECKED_DIV(L, R) (((R) != 0) ? ((L) / (R)) : (flow_div_by_zero_handler(), (L) * 0))
#endif
#ifndef FLOW_CHECKED_MOD
#define FLOW_CHECKED_MOD(L, R) (((R) != 0) ? ((L) % (R)) : (flow_div_by_zero_handler(), (L) * 0))
#endif
#ifndef FLOW_CHECKED_SHL
#define FLOW_CHECKED_SHL(L, R) ((((R) >= 0) && ((unsigned long long)(R) < (sizeof(L) * 8ull)) && ((L) >= 0)) ? ((L) << (R)) : (flow_shift_ub_handler(), (L) * 0))
#endif
#ifndef FLOW_CHECKED_SHR
#define FLOW_CHECKED_SHR(L, R) ((((R) >= 0) && ((unsigned long long)(R) < (sizeof(L) * 8ull))) ? ((L) >> (R)) : (flow_shift_ub_handler(), (L) * 0))
#endif
#include <math.h>
void* _ui_state = NULL;
static inline float i32_to_f32(int32_t v) { return (float)v; }
/* Host stub for @gpu kernels (device codegen replaces this). */
static inline int32_t gpu_thread_id(void) { return 0; }
int64_t floordiv_i64_i64(int64_t a, int64_t b);
int64_t f_gauss_i64_i64(int64_t a0, int64_t b0);
__int128 count_rect_i64_i64_i64_i64(int64_t x0, int64_t x1, int64_t y0, int64_t y1);
int64_t count_parity_i64_i64_i64(int64_t lo, int64_t hi, int64_t parity);
__int128 count_diamond_i64_i64_i64_i64(int64_t u0, int64_t u1, int64_t v0, int64_t v1);
int64_t ceil_div_i64_i64(int64_t n, int64_t d);
int64_t hkey_i64_i64_i64_i64_i64(int64_t a0, int64_t a1, int64_t b0, int64_t b1, int64_t tag);
int64_t hslot_i64_ptr_i64_ptr_i8(int64_t key, int64_t* keys, int8_t* used);
int64_t brute_rect_i64_i64_i64_i64(int64_t x0, int64_t x1, int64_t y0, int64_t y1);
int64_t brute_diamond_i64_i64_i64_i64(int64_t u0, int64_t u1, int64_t v0, int64_t v1);
void rect_to_diamond_i64_i64_i64_i64_i64_ptr_i64_ptr_i64_ptr_i64_ptr_i64(int64_t x0, int64_t x1, int64_t y0, int64_t y1, int64_t r, int64_t* u0, int64_t* u1, int64_t* v0, int64_t* v1);
void diamond_to_rect_i64_i64_i64_i64_i64_ptr_i64_ptr_i64_ptr_i64_ptr_i64(int64_t u0, int64_t u1, int64_t v0, int64_t v1, int64_t r, int64_t* x0, int64_t* x1, int64_t* y0, int64_t* y1);
int64_t sum_region_i32_i64_i64_i64_i64_ptr_i64_ptr_i64_ptr_i8_ptr_i64(int32_t tag, int64_t a0, int64_t a1, int64_t b0, int64_t b1, int64_t* keys, int64_t* vals, int8_t* used, int64_t* buf);
int32_t main(void);
static const int64_t MOD = 1000000007;
static const int64_t LIM = 1000000000000000;
static const int64_t BRUTE = 4000;
static const int64_t CAP = 1048576;
int64_t floordiv_i64_i64(int64_t a, int64_t b) {
if (b == 0) {
return 0;
}
int64_t q = FLOW_CHECKED_DIV((a), (b));
int64_t r = (a - (q * b));
if ((r != 0 && r > 0 != b > 0)) {
q = (q - 1);
}
return q;
}
int64_t f_gauss_i64_i64(int64_t a0, int64_t b0) {
int64_t a = a0;
int64_t b = b0;
int64_t cnt = 0;
while ((a != 0 || b != 0)) {
if (((a ^ b) & 1) != 0) {
a = (a - 1);
cnt = (cnt + 1);
}
int64_t na = floordiv_i64_i64((b - a), 2);
int64_t nb = floordiv_i64_i64((-(a + b)), 2);
a = na;
b = nb;
}
return cnt;
}
__int128 count_rect_i64_i64_i64_i64(int64_t x0, int64_t x1, int64_t y0, int64_t y1) {
if ((x0 > x1 || y0 > y1)) {
return ((__int128)(0));
}
__int128 w = ((__int128)(((x1 - x0) + 1)));
__int128 h = ((__int128)(((y1 - y0) + 1)));
return (w * h);
}
int64_t count_parity_i64_i64_i64(int64_t lo, int64_t hi, int64_t parity) {
if (lo > hi) {
return 0;
}
int64_t first = lo;
if ((lo & 1) != parity) {
first = (lo + 1);
}
if (first > hi) {
return 0;
}
return (floordiv_i64_i64((hi - first), 2) + 1);
}
__int128 count_diamond_i64_i64_i64_i64(int64_t u0, int64_t u1, int64_t v0, int64_t v1) {
if ((u0 > u1 || v0 > v1)) {
return ((__int128)(0));
}
__int128 eu = ((__int128)(count_parity_i64_i64_i64(u0, u1, 0)));
__int128 ou = (((__int128)(((u1 - u0) + 1))) - eu);
__int128 ev = ((__int128)(count_parity_i64_i64_i64(v0, v1, 0)));
__int128 ov = (((__int128)(((v1 - v0) + 1))) - ev);
return ((eu * ev) + (ou * ov));
}
int64_t ceil_div_i64_i64(int64_t n, int64_t d) {
return (-floordiv_i64_i64((-n), d));
}
int64_t hkey_i64_i64_i64_i64_i64(int64_t a0, int64_t a1, int64_t b0, int64_t b1, int64_t tag) {
__int128 P = ((__int128)(1000000000000000003));
__int128 k = ((__int128)(tag));
k = FLOW_CHECKED_MOD((((k * ((__int128)(1000003))) + ((__int128)(a0)))), (P));
k = FLOW_CHECKED_MOD((((k * ((__int128)(1000003))) + ((__int128)(a1)))), (P));
k = FLOW_CHECKED_MOD((((k * ((__int128)(1000003))) + ((__int128)(b0)))), (P));
k = FLOW_CHECKED_MOD((((k * ((__int128)(1000003))) + ((__int128)(b1)))), (P));
if (k < 0) {
k = (k + P);
}
return ((int64_t)(k));
}
int64_t hslot_i64_ptr_i64_ptr_i8(int64_t key, int64_t* keys, int8_t* used) {
int64_t h = FLOW_CHECKED_MOD((key), (CAP));
if (h < 0) {
h = (h + CAP);
}
while ((used[h] == 1 && keys[h] != key)) {
h = (h + 1);
if (h == CAP) {
h = 0;
}
}
return h;
}
int64_t brute_rect_i64_i64_i64_i64(int64_t x0, int64_t x1, int64_t y0, int64_t y1) {
int64_t s = 0;
int64_t a = x0;
while (a <= x1) {
int64_t b = y0;
while (b <= y1) {
s = (s + f_gauss_i64_i64(a, b));
b = (b + 1);
}
a = (a + 1);
}
return FLOW_CHECKED_MOD((s), (MOD));
}
int64_t brute_diamond_i64_i64_i64_i64(int64_t u0, int64_t u1, int64_t v0, int64_t v1) {
int64_t s = 0;
int64_t u = u0;
while (u <= u1) {
int64_t parity = (u & 1);
int64_t v = v0;
if ((v & 1) != parity) {
v = (v + 1);
}
while (v <= v1) {
s = (s + f_gauss_i64_i64(floordiv_i64_i64((u + v), 2), floordiv_i64_i64((u - v), 2)));
v = (v + 2);
}
u = (u + 1);
}
return FLOW_CHECKED_MOD((s), (MOD));
}
void rect_to_diamond_i64_i64_i64_i64_i64_ptr_i64_ptr_i64_ptr_i64_ptr_i64(int64_t x0, int64_t x1, int64_t y0, int64_t y1, int64_t r, int64_t* u0, int64_t* u1, int64_t* v0, int64_t* v1) {
u0[0] = ((-x1) + r);
u1[0] = ((-x0) + r);
v0[0] = y0;
v1[0] = y1;
}
void diamond_to_rect_i64_i64_i64_i64_i64_ptr_i64_ptr_i64_ptr_i64_ptr_i64(int64_t u0, int64_t u1, int64_t v0, int64_t v1, int64_t r, int64_t* x0, int64_t* x1, int64_t* y0, int64_t* y1) {
x0[0] = ceil_div_i64_i64((r - v1), 2);
x1[0] = floordiv_i64_i64((r - v0), 2);
y0[0] = ceil_div_i64_i64((r - u1), 2);
y1[0] = floordiv_i64_i64((r - u0), 2);
}
int64_t sum_region_i32_i64_i64_i64_i64_ptr_i64_ptr_i64_ptr_i8_ptr_i64(int32_t tag, int64_t a0, int64_t a1, int64_t b0, int64_t b1, int64_t* keys, int64_t* vals, int8_t* used, int64_t* buf) {
if ((a0 > a1 || b0 > b1)) {
return 0;
}
if (tag == 0) {
__int128 n0 = count_rect_i64_i64_i64_i64(a0, a1, b0, b1);
if (n0 <= ((__int128)(BRUTE))) {
return brute_rect_i64_i64_i64_i64(a0, a1, b0, b1);
}
} else {
__int128 n1 = count_diamond_i64_i64_i64_i64(a0, a1, b0, b1);
if (n1 <= ((__int128)(BRUTE))) {
return brute_diamond_i64_i64_i64_i64(a0, a1, b0, b1);
}
}
int64_t key = hkey_i64_i64_i64_i64_i64(a0, a1, b0, b1, ((int64_t)(tag)));
int64_t slot = hslot_i64_ptr_i64_ptr_i8(key, keys, used);
if ((used[slot] == 1 && keys[slot] == key)) {
return vals[slot];
}
int64_t res = 0;
if (tag == 0) {
rect_to_diamond_i64_i64_i64_i64_i64_ptr_i64_ptr_i64_ptr_i64_ptr_i64(a0, a1, b0, b1, 0, buf, (buf + 1), (buf + 2), (buf + 3));
res = sum_region_i32_i64_i64_i64_i64_ptr_i64_ptr_i64_ptr_i8_ptr_i64(1, buf[0], buf[1], buf[2], buf[3], keys, vals, used, buf);
rect_to_diamond_i64_i64_i64_i64_i64_ptr_i64_ptr_i64_ptr_i64_ptr_i64(a0, a1, b0, b1, 1, buf, (buf + 1), (buf + 2), (buf + 3));
int64_t du0 = buf[0];
int64_t du1 = buf[1];
int64_t dv0 = buf[2];
int64_t dv1 = buf[3];
res = FLOW_CHECKED_MOD(((res + sum_region_i32_i64_i64_i64_i64_ptr_i64_ptr_i64_ptr_i8_ptr_i64(1, du0, du1, dv0, dv1, keys, vals, used, buf))), (MOD));
res = FLOW_CHECKED_MOD(((res + ((int64_t)(FLOW_CHECKED_MOD((count_diamond_i64_i64_i64_i64(du0, du1, dv0, dv1)), (((__int128)(MOD)))))))), (MOD));
} else {
diamond_to_rect_i64_i64_i64_i64_i64_ptr_i64_ptr_i64_ptr_i64_ptr_i64(a0, a1, b0, b1, 0, buf, (buf + 1), (buf + 2), (buf + 3));
res = sum_region_i32_i64_i64_i64_i64_ptr_i64_ptr_i64_ptr_i8_ptr_i64(0, buf[0], buf[1], buf[2], buf[3], keys, vals, used, buf);
diamond_to_rect_i64_i64_i64_i64_i64_ptr_i64_ptr_i64_ptr_i64_ptr_i64(a0, a1, b0, b1, 1, buf, (buf + 1), (buf + 2), (buf + 3));
int64_t rx0 = buf[0];
int64_t rx1 = buf[1];
int64_t ry0 = buf[2];
int64_t ry1 = buf[3];
res = FLOW_CHECKED_MOD(((res + sum_region_i32_i64_i64_i64_i64_ptr_i64_ptr_i64_ptr_i8_ptr_i64(0, rx0, rx1, ry0, ry1, keys, vals, used, buf))), (MOD));
res = FLOW_CHECKED_MOD(((res + ((int64_t)(FLOW_CHECKED_MOD((count_rect_i64_i64_i64_i64(rx0, rx1, ry0, ry1)), (((__int128)(MOD)))))))), (MOD));
}
int64_t store = hslot_i64_ptr_i64_ptr_i8(key, keys, used);
used[store] = 1;
keys[store] = key;
vals[store] = res;
return res;
}
int32_t main(void) {
int64_t* keys = (int64_t*)(calloc(CAP, 8));
int64_t* vals = (int64_t*)(calloc(CAP, 8));
int8_t* used = (int8_t*)(calloc(CAP, 1));
int64_t* buf = (int64_t*)(calloc(4, 8));
if ((((keys == NULL || vals == NULL) || used == NULL) || buf == NULL)) {
return 1;
}
int64_t ans = sum_region_i32_i64_i64_i64_i64_ptr_i64_ptr_i64_ptr_i8_ptr_i64(0, (-LIM), LIM, (-LIM), LIM, keys, vals, used, buf);
printf("%lld\n", ans);
free(keys);
free(vals);
free(used);
free(buf);
return 0;
}
Generated MLIR
module {
llvm.func @printf(!llvm.ptr, ...) -> i32
llvm.mlir.global internal constant @str_0("%lld\n\00") {addr_space = 0 : i32} : !llvm.array<6 x i8>
// Constant: MOD
llvm.mlir.global internal constant @MOD(1000000007 : i64) : i64
// Constant: LIM
llvm.mlir.global internal constant @LIM(1000000000000000 : i64) : i64
// Constant: BRUTE
llvm.mlir.global internal constant @BRUTE(4000 : i64) : i64
// Constant: CAP
llvm.mlir.global internal constant @CAP(1048576 : i64) : i64
func.func private @calloc(i64, i64) -> !llvm.ptr
func.func private @free(!llvm.ptr) -> ()
func.func @floordiv(%arg0: i64, %arg1: i64) -> i64 {
%0 = arith.constant 0 : i32
%2 = arith.extsi %0 : i32 to i64
%1 = arith.cmpi eq, %arg1, %2 : i64
cf.cond_br %1, ^bb0, ^bb1
^bb0:
%3 = arith.constant 0 : i32
%4 = arith.extsi %3 : i32 to i64
func.return %4 : i64
^bb1:
cf.br ^bb2
^bb2:
%5 = arith.divsi %arg0, %arg1 : i64
%6 = arith.muli %5, %arg1 : i64
%7 = arith.subi %arg0, %6 : i64
%8 = arith.constant 0 : i32
%10 = arith.extsi %8 : i32 to i64
%9 = arith.cmpi ne, %7, %10 : i64
%11 = scf.if %9 -> (i1) {
%12 = arith.constant 0 : i32
%14 = arith.extsi %12 : i32 to i64
%13 = arith.cmpi sgt, %7, %14 : i64
%15 = arith.constant 0 : i32
%17 = arith.extsi %15 : i32 to i64
%16 = arith.cmpi sgt, %arg1, %17 : i64
%18 = arith.cmpi ne, %13, %16 : i1
scf.yield %18 : i1
} else {
%19 = arith.constant false
scf.yield %19 : i1
}
%20 = scf.if %11 -> (i64) {
%21 = arith.constant 1 : i32
%23 = arith.extsi %21 : i32 to i64
%22 = arith.subi %5, %23 : i64
scf.yield %22 : i64
} else {
scf.yield %5 : i64
}
func.return %20 : i64
}
func.func @f_gauss(%arg0: i64, %arg1: i64) -> i64 {
%24 = llvm.mlir.constant(1 : i64) : i64
%25 = llvm.alloca %24 x i64 : (i64) -> !llvm.ptr
llvm.store %arg0, %25 : i64, !llvm.ptr
%26 = llvm.mlir.constant(1 : i64) : i64
%27 = llvm.alloca %26 x i64 : (i64) -> !llvm.ptr
llvm.store %arg1, %27 : i64, !llvm.ptr
%28 = arith.constant 0 : i32
%29 = arith.extsi %28 : i32 to i64
%30 = llvm.mlir.constant(1 : i64) : i64
%31 = llvm.alloca %30 x i64 : (i64) -> !llvm.ptr
llvm.store %29, %31 : i64, !llvm.ptr
cf.br ^bb3
^bb3:
%32 = llvm.load %25 : !llvm.ptr -> i64
%33 = arith.constant 0 : i32
%35 = arith.extsi %33 : i32 to i64
%34 = arith.cmpi ne, %32, %35 : i64
%36 = scf.if %34 -> (i1) {
%37 = arith.constant true
scf.yield %37 : i1
} else {
%38 = llvm.load %27 : !llvm.ptr -> i64
%39 = arith.constant 0 : i32
%41 = arith.extsi %39 : i32 to i64
%40 = arith.cmpi ne, %38, %41 : i64
scf.yield %40 : i1
}
cf.cond_br %36, ^bb4, ^bb5
^bb4:
%42 = llvm.load %25 : !llvm.ptr -> i64
%43 = llvm.load %27 : !llvm.ptr -> i64
%44 = arith.xori %42, %43 : i64
%45 = arith.constant 1 : i32
%47 = arith.extsi %45 : i32 to i64
%46 = arith.andi %44, %47 : i64
%48 = arith.constant 0 : i32
%50 = arith.extsi %48 : i32 to i64
%49 = arith.cmpi ne, %46, %50 : i64
cf.cond_br %49, ^bb6, ^bb7
^bb6:
%51 = llvm.load %25 : !llvm.ptr -> i64
%52 = arith.constant 1 : i32
%54 = arith.extsi %52 : i32 to i64
%53 = arith.subi %51, %54 : i64
llvm.store %53, %25 : i64, !llvm.ptr
%55 = llvm.load %31 : !llvm.ptr -> i64
%56 = arith.constant 1 : i32
%58 = arith.extsi %56 : i32 to i64
%57 = arith.addi %55, %58 : i64
llvm.store %57, %31 : i64, !llvm.ptr
cf.br ^bb8
^bb7:
cf.br ^bb8
^bb8:
%60 = llvm.load %27 : !llvm.ptr -> i64
%61 = llvm.load %25 : !llvm.ptr -> i64
%62 = arith.subi %60, %61 : i64
%63 = arith.constant 2 : i32
%64 = arith.extsi %63 : i32 to i64
%59 = func.call @floordiv(%62, %64) : (i64, i64) -> i64
%66 = llvm.load %25 : !llvm.ptr -> i64
%67 = llvm.load %27 : !llvm.ptr -> i64
%68 = arith.addi %66, %67 : i64
%70 = arith.constant 0 : i64
%69 = arith.subi %70, %68 : i64
%71 = arith.constant 2 : i32
%72 = arith.extsi %71 : i32 to i64
%65 = func.call @floordiv(%69, %72) : (i64, i64) -> i64
llvm.store %59, %25 : i64, !llvm.ptr
llvm.store %65, %27 : i64, !llvm.ptr
cf.br ^bb3
^bb5:
%73 = llvm.load %31 : !llvm.ptr -> i64
func.return %73 : i64
}
func.func @count_rect(%arg0: i64, %arg1: i64, %arg2: i64, %arg3: i64) -> i128 {
%74 = arith.cmpi sgt, %arg0, %arg1 : i64
%75 = scf.if %74 -> (i1) {
%76 = arith.constant true
scf.yield %76 : i1
} else {
%77 = arith.cmpi sgt, %arg2, %arg3 : i64
scf.yield %77 : i1
}
cf.cond_br %75, ^bb9, ^bb10
^bb9:
%78 = arith.constant 0 : i32
%79 = arith.extsi %78 : i32 to i128
func.return %79 : i128
^bb10:
cf.br ^bb11
^bb11:
%80 = arith.subi %arg1, %arg0 : i64
%81 = arith.constant 1 : i32
%83 = arith.extsi %81 : i32 to i64
%82 = arith.addi %80, %83 : i64
%84 = arith.extsi %82 : i64 to i128
%85 = arith.subi %arg3, %arg2 : i64
%86 = arith.constant 1 : i32
%88 = arith.extsi %86 : i32 to i64
%87 = arith.addi %85, %88 : i64
%89 = arith.extsi %87 : i64 to i128
%91 = arith.trunci %84 : i128 to i64
%92 = arith.trunci %89 : i128 to i64
%90 = arith.muli %91, %92 : i64
%93 = arith.extsi %90 : i64 to i128
func.return %93 : i128
}
func.func @count_parity(%arg0: i64, %arg1: i64, %arg2: i64) -> i64 {
%94 = arith.cmpi sgt, %arg0, %arg1 : i64
cf.cond_br %94, ^bb12, ^bb13
^bb12:
%95 = arith.constant 0 : i32
%96 = arith.extsi %95 : i32 to i64
func.return %96 : i64
^bb13:
cf.br ^bb14
^bb14:
%97 = llvm.mlir.constant(1 : i64) : i64
%98 = llvm.alloca %97 x i64 : (i64) -> !llvm.ptr
llvm.store %arg0, %98 : i64, !llvm.ptr
%99 = arith.constant 1 : i32
%101 = arith.extsi %99 : i32 to i64
%100 = arith.andi %arg0, %101 : i64
%102 = arith.cmpi ne, %100, %arg2 : i64
cf.cond_br %102, ^bb15, ^bb16
^bb15:
%103 = arith.constant 1 : i32
%105 = arith.extsi %103 : i32 to i64
%104 = arith.addi %arg0, %105 : i64
llvm.store %104, %98 : i64, !llvm.ptr
cf.br ^bb17
^bb16:
cf.br ^bb17
^bb17:
%106 = llvm.load %98 : !llvm.ptr -> i64
%107 = arith.cmpi sgt, %106, %arg1 : i64
cf.cond_br %107, ^bb18, ^bb19
^bb18:
%108 = arith.constant 0 : i32
%109 = arith.extsi %108 : i32 to i64
func.return %109 : i64
^bb19:
cf.br ^bb20
^bb20:
%111 = llvm.load %98 : !llvm.ptr -> i64
%112 = arith.subi %arg1, %111 : i64
%113 = arith.constant 2 : i32
%114 = arith.extsi %113 : i32 to i64
%110 = func.call @floordiv(%112, %114) : (i64, i64) -> i64
%115 = arith.constant 1 : i32
%117 = arith.extsi %115 : i32 to i64
%116 = arith.addi %110, %117 : i64
func.return %116 : i64
}
func.func @count_diamond(%arg0: i64, %arg1: i64, %arg2: i64, %arg3: i64) -> i128 {
%118 = arith.cmpi sgt, %arg0, %arg1 : i64
%119 = scf.if %118 -> (i1) {
%120 = arith.constant true
scf.yield %120 : i1
} else {
%121 = arith.cmpi sgt, %arg2, %arg3 : i64
scf.yield %121 : i1
}
cf.cond_br %119, ^bb21, ^bb22
^bb21:
%122 = arith.constant 0 : i32
%123 = arith.extsi %122 : i32 to i128
func.return %123 : i128
^bb22:
cf.br ^bb23
^bb23:
%125 = arith.constant 0 : i32
%126 = arith.extsi %125 : i32 to i64
%124 = func.call @count_parity(%arg0, %arg1, %126) : (i64, i64, i64) -> i64
%127 = arith.extsi %124 : i64 to i128
%128 = arith.subi %arg1, %arg0 : i64
%129 = arith.constant 1 : i32
%131 = arith.extsi %129 : i32 to i64
%130 = arith.addi %128, %131 : i64
%132 = arith.extsi %130 : i64 to i128
%134 = arith.trunci %132 : i128 to i64
%135 = arith.trunci %127 : i128 to i64
%133 = arith.subi %134, %135 : i64
%136 = arith.extsi %133 : i64 to i128
%138 = arith.constant 0 : i32
%139 = arith.extsi %138 : i32 to i64
%137 = func.call @count_parity(%arg2, %arg3, %139) : (i64, i64, i64) -> i64
%140 = arith.extsi %137 : i64 to i128
%141 = arith.subi %arg3, %arg2 : i64
%142 = arith.constant 1 : i32
%144 = arith.extsi %142 : i32 to i64
%143 = arith.addi %141, %144 : i64
%145 = arith.extsi %143 : i64 to i128
%147 = arith.trunci %145 : i128 to i64
%148 = arith.trunci %140 : i128 to i64
%146 = arith.subi %147, %148 : i64
%149 = arith.extsi %146 : i64 to i128
%151 = arith.trunci %127 : i128 to i64
%152 = arith.trunci %140 : i128 to i64
%150 = arith.muli %151, %152 : i64
%154 = arith.trunci %136 : i128 to i64
%155 = arith.trunci %149 : i128 to i64
%153 = arith.muli %154, %155 : i64
%156 = arith.addi %150, %153 : i64
%157 = arith.extsi %156 : i64 to i128
func.return %157 : i128
}
func.func @ceil_div(%arg0: i64, %arg1: i64) -> i64 {
%160 = arith.constant 0 : i64
%159 = arith.subi %160, %arg0 : i64
%158 = func.call @floordiv(%159, %arg1) : (i64, i64) -> i64
%162 = arith.constant 0 : i64
%161 = arith.subi %162, %158 : i64
func.return %161 : i64
}
func.func @hkey(%arg0: i64, %arg1: i64, %arg2: i64, %arg3: i64, %arg4: i64) -> i64 {
%163 = arith.constant 999999995705032707 : i32
%164 = arith.extsi %163 : i32 to i128
%165 = arith.extsi %arg4 : i64 to i128
%166 = llvm.mlir.constant(1 : i64) : i64
%167 = llvm.alloca %166 x i128 : (i64) -> !llvm.ptr
llvm.store %165, %167 : i128, !llvm.ptr
%168 = llvm.load %167 : !llvm.ptr -> i128
%169 = arith.constant 1000003 : i32
%170 = arith.extsi %169 : i32 to i128
%172 = arith.trunci %168 : i128 to i64
%173 = arith.trunci %170 : i128 to i64
%171 = arith.muli %172, %173 : i64
%174 = arith.extsi %arg0 : i64 to i128
%176 = arith.trunci %174 : i128 to i64
%175 = arith.addi %171, %176 : i64
%178 = arith.trunci %164 : i128 to i64
%177 = arith.remsi %175, %178 : i64
%179 = arith.extsi %177 : i64 to i128
llvm.store %179, %167 : i128, !llvm.ptr
%180 = llvm.load %167 : !llvm.ptr -> i128
%181 = arith.constant 1000003 : i32
%182 = arith.extsi %181 : i32 to i128
%184 = arith.trunci %180 : i128 to i64
%185 = arith.trunci %182 : i128 to i64
%183 = arith.muli %184, %185 : i64
%186 = arith.extsi %arg1 : i64 to i128
%188 = arith.trunci %186 : i128 to i64
%187 = arith.addi %183, %188 : i64
%190 = arith.trunci %164 : i128 to i64
%189 = arith.remsi %187, %190 : i64
%191 = arith.extsi %189 : i64 to i128
llvm.store %191, %167 : i128, !llvm.ptr
%192 = llvm.load %167 : !llvm.ptr -> i128
%193 = arith.constant 1000003 : i32
%194 = arith.extsi %193 : i32 to i128
%196 = arith.trunci %192 : i128 to i64
%197 = arith.trunci %194 : i128 to i64
%195 = arith.muli %196, %197 : i64
%198 = arith.extsi %arg2 : i64 to i128
%200 = arith.trunci %198 : i128 to i64
%199 = arith.addi %195, %200 : i64
%202 = arith.trunci %164 : i128 to i64
%201 = arith.remsi %199, %202 : i64
%203 = arith.extsi %201 : i64 to i128
llvm.store %203, %167 : i128, !llvm.ptr
%204 = llvm.load %167 : !llvm.ptr -> i128
%205 = arith.constant 1000003 : i32
%206 = arith.extsi %205 : i32 to i128
%208 = arith.trunci %204 : i128 to i64
%209 = arith.trunci %206 : i128 to i64
%207 = arith.muli %208, %209 : i64
%210 = arith.extsi %arg3 : i64 to i128
%212 = arith.trunci %210 : i128 to i64
%211 = arith.addi %207, %212 : i64
%214 = arith.trunci %164 : i128 to i64
%213 = arith.remsi %211, %214 : i64
%215 = arith.extsi %213 : i64 to i128
llvm.store %215, %167 : i128, !llvm.ptr
%216 = llvm.load %167 : !llvm.ptr -> i128
%217 = arith.constant 0 : i32
%219 = arith.trunci %216 : i128 to i64
%220 = arith.extsi %217 : i32 to i64
%218 = arith.cmpi slt, %219, %220 : i64
cf.cond_br %218, ^bb24, ^bb25
^bb24:
%221 = llvm.load %167 : !llvm.ptr -> i128
%223 = arith.trunci %221 : i128 to i64
%224 = arith.trunci %164 : i128 to i64
%222 = arith.addi %223, %224 : i64
%225 = arith.extsi %222 : i64 to i128
llvm.store %225, %167 : i128, !llvm.ptr
cf.br ^bb26
^bb25:
cf.br ^bb26
^bb26:
%226 = llvm.load %167 : !llvm.ptr -> i128
%227 = arith.trunci %226 : i128 to i64
func.return %227 : i64
}
func.func @hslot(%arg0: i64, %arg1: !llvm.ptr, %arg2: !llvm.ptr) -> i64 {
%228 = llvm.mlir.addressof @CAP : !llvm.ptr
%229 = llvm.load %228 : !llvm.ptr -> i64
%230 = arith.remsi %arg0, %229 : i64
%231 = llvm.mlir.constant(1 : i64) : i64
%232 = llvm.alloca %231 x i64 : (i64) -> !llvm.ptr
llvm.store %230, %232 : i64, !llvm.ptr
%233 = llvm.load %232 : !llvm.ptr -> i64
%234 = arith.constant 0 : i32
%236 = arith.extsi %234 : i32 to i64
%235 = arith.cmpi slt, %233, %236 : i64
cf.cond_br %235, ^bb27, ^bb28
^bb27:
%237 = llvm.load %232 : !llvm.ptr -> i64
%238 = llvm.mlir.addressof @CAP : !llvm.ptr
%239 = llvm.load %238 : !llvm.ptr -> i64
%240 = arith.addi %237, %239 : i64
llvm.store %240, %232 : i64, !llvm.ptr
cf.br ^bb29
^bb28:
cf.br ^bb29
^bb29:
cf.br ^bb30
^bb30:
%242 = llvm.load %232 : !llvm.ptr -> i64
%243 = llvm.getelementptr %arg2[%242] : (!llvm.ptr, i64) -> !llvm.ptr, i8
%241 = llvm.load %243 : !llvm.ptr -> i8
%244 = arith.constant 1 : i32
%246 = arith.extsi %241 : i8 to i32
%245 = arith.cmpi eq, %246, %244 : i32
%247 = scf.if %245 -> (i1) {
%249 = llvm.load %232 : !llvm.ptr -> i64
%250 = llvm.getelementptr %arg1[%249] : (!llvm.ptr, i64) -> !llvm.ptr, i64
%248 = llvm.load %250 : !llvm.ptr -> i64
%251 = arith.cmpi ne, %248, %arg0 : i64
scf.yield %251 : i1
} else {
%252 = arith.constant false
scf.yield %252 : i1
}
cf.cond_br %247, ^bb31, ^bb32
^bb31:
%253 = llvm.load %232 : !llvm.ptr -> i64
%254 = arith.constant 1 : i32
%256 = arith.extsi %254 : i32 to i64
%255 = arith.addi %253, %256 : i64
llvm.store %255, %232 : i64, !llvm.ptr
%257 = llvm.load %232 : !llvm.ptr -> i64
%258 = llvm.mlir.addressof @CAP : !llvm.ptr
%259 = llvm.load %258 : !llvm.ptr -> i64
%260 = arith.cmpi eq, %257, %259 : i64
cf.cond_br %260, ^bb33, ^bb34
^bb33:
%261 = arith.constant 0 : i32
%262 = arith.extsi %261 : i32 to i64
llvm.store %262, %232 : i64, !llvm.ptr
cf.br ^bb35
^bb34:
cf.br ^bb35
^bb35:
cf.br ^bb30
^bb32:
%263 = llvm.load %232 : !llvm.ptr -> i64
func.return %263 : i64
}
func.func @brute_rect(%arg0: i64, %arg1: i64, %arg2: i64, %arg3: i64) -> i64 {
%264 = arith.constant 0 : i32
%265 = arith.extsi %264 : i32 to i64
%266 = llvm.mlir.constant(1 : i64) : i64
%267 = llvm.alloca %266 x i64 : (i64) -> !llvm.ptr
llvm.store %265, %267 : i64, !llvm.ptr
%268 = llvm.mlir.constant(1 : i64) : i64
%269 = llvm.alloca %268 x i64 : (i64) -> !llvm.ptr
llvm.store %arg0, %269 : i64, !llvm.ptr
cf.br ^bb36
^bb36:
%270 = llvm.load %269 : !llvm.ptr -> i64
%271 = arith.cmpi sle, %270, %arg1 : i64
cf.cond_br %271, ^bb37, ^bb38
^bb37:
%272 = llvm.mlir.constant(1 : i64) : i64
%273 = llvm.alloca %272 x i64 : (i64) -> !llvm.ptr
llvm.store %arg2, %273 : i64, !llvm.ptr
cf.br ^bb39
^bb39:
%274 = llvm.load %273 : !llvm.ptr -> i64
%275 = arith.cmpi sle, %274, %arg3 : i64
cf.cond_br %275, ^bb40, ^bb41
^bb40:
%276 = llvm.load %267 : !llvm.ptr -> i64
%278 = llvm.load %269 : !llvm.ptr -> i64
%279 = llvm.load %273 : !llvm.ptr -> i64
%277 = func.call @f_gauss(%278, %279) : (i64, i64) -> i64
%280 = arith.addi %276, %277 : i64
llvm.store %280, %267 : i64, !llvm.ptr
%281 = llvm.load %273 : !llvm.ptr -> i64
%282 = arith.constant 1 : i32
%284 = arith.extsi %282 : i32 to i64
%283 = arith.addi %281, %284 : i64
llvm.store %283, %273 : i64, !llvm.ptr
cf.br ^bb39
^bb41:
%285 = llvm.load %269 : !llvm.ptr -> i64
%286 = arith.constant 1 : i32
%288 = arith.extsi %286 : i32 to i64
%287 = arith.addi %285, %288 : i64
llvm.store %287, %269 : i64, !llvm.ptr
cf.br ^bb36
^bb38:
%289 = llvm.load %267 : !llvm.ptr -> i64
%290 = llvm.mlir.addressof @MOD : !llvm.ptr
%291 = llvm.load %290 : !llvm.ptr -> i64
%292 = arith.remsi %289, %291 : i64
func.return %292 : i64
}
func.func @brute_diamond(%arg0: i64, %arg1: i64, %arg2: i64, %arg3: i64) -> i64 {
%293 = arith.constant 0 : i32
%294 = arith.extsi %293 : i32 to i64
%295 = llvm.mlir.constant(1 : i64) : i64
%296 = llvm.alloca %295 x i64 : (i64) -> !llvm.ptr
llvm.store %294, %296 : i64, !llvm.ptr
%297 = llvm.mlir.constant(1 : i64) : i64
%298 = llvm.alloca %297 x i64 : (i64) -> !llvm.ptr
llvm.store %arg0, %298 : i64, !llvm.ptr
cf.br ^bb42
^bb42:
%299 = llvm.load %298 : !llvm.ptr -> i64
%300 = arith.cmpi sle, %299, %arg1 : i64
cf.cond_br %300, ^bb43, ^bb44
^bb43:
%301 = llvm.load %298 : !llvm.ptr -> i64
%302 = arith.constant 1 : i32
%304 = arith.extsi %302 : i32 to i64
%303 = arith.andi %301, %304 : i64
%305 = llvm.mlir.constant(1 : i64) : i64
%306 = llvm.alloca %305 x i64 : (i64) -> !llvm.ptr
llvm.store %arg2, %306 : i64, !llvm.ptr
%307 = llvm.load %306 : !llvm.ptr -> i64
%308 = arith.constant 1 : i32
%310 = arith.extsi %308 : i32 to i64
%309 = arith.andi %307, %310 : i64
%311 = arith.cmpi ne, %309, %303 : i64
cf.cond_br %311, ^bb45, ^bb46
^bb45:
%312 = llvm.load %306 : !llvm.ptr -> i64
%313 = arith.constant 1 : i32
%315 = arith.extsi %313 : i32 to i64
%314 = arith.addi %312, %315 : i64
llvm.store %314, %306 : i64, !llvm.ptr
cf.br ^bb47
^bb46:
cf.br ^bb47
^bb47:
cf.br ^bb48
^bb48:
%316 = llvm.load %306 : !llvm.ptr -> i64
%317 = arith.cmpi sle, %316, %arg3 : i64
cf.cond_br %317, ^bb49, ^bb50
^bb49:
%318 = llvm.load %296 : !llvm.ptr -> i64
%321 = llvm.load %298 : !llvm.ptr -> i64
%322 = llvm.load %306 : !llvm.ptr -> i64
%323 = arith.addi %321, %322 : i64
%324 = arith.constant 2 : i32
%325 = arith.extsi %324 : i32 to i64
%320 = func.call @floordiv(%323, %325) : (i64, i64) -> i64
%327 = llvm.load %298 : !llvm.ptr -> i64
%328 = llvm.load %306 : !llvm.ptr -> i64
%329 = arith.subi %327, %328 : i64
%330 = arith.constant 2 : i32
%331 = arith.extsi %330 : i32 to i64
%326 = func.call @floordiv(%329, %331) : (i64, i64) -> i64
%319 = func.call @f_gauss(%320, %326) : (i64, i64) -> i64
%332 = arith.addi %318, %319 : i64
llvm.store %332, %296 : i64, !llvm.ptr
%333 = llvm.load %306 : !llvm.ptr -> i64
%334 = arith.constant 2 : i32
%336 = arith.extsi %334 : i32 to i64
%335 = arith.addi %333, %336 : i64
llvm.store %335, %306 : i64, !llvm.ptr
cf.br ^bb48
^bb50:
%337 = llvm.load %298 : !llvm.ptr -> i64
%338 = arith.constant 1 : i32
%340 = arith.extsi %338 : i32 to i64
%339 = arith.addi %337, %340 : i64
llvm.store %339, %298 : i64, !llvm.ptr
cf.br ^bb42
^bb44:
%341 = llvm.load %296 : !llvm.ptr -> i64
%342 = llvm.mlir.addressof @MOD : !llvm.ptr
%343 = llvm.load %342 : !llvm.ptr -> i64
%344 = arith.remsi %341, %343 : i64
func.return %344 : i64
}
func.func @rect_to_diamond(%arg0: i64, %arg1: i64, %arg2: i64, %arg3: i64, %arg4: i64, %arg5: !llvm.ptr, %arg6: !llvm.ptr, %arg7: !llvm.ptr, %arg8: !llvm.ptr) -> () {
%346 = arith.constant 0 : i64
%345 = arith.subi %346, %arg1 : i64
%347 = arith.addi %345, %arg4 : i64
%348 = arith.constant 0 : i32
%349 = arith.extsi %348 : i32 to i64
%350 = llvm.getelementptr %arg5[%349] : (!llvm.ptr, i64) -> !llvm.ptr, i64
llvm.store %347, %350 : i64, !llvm.ptr
%352 = arith.constant 0 : i64
%351 = arith.subi %352, %arg0 : i64
%353 = arith.addi %351, %arg4 : i64
%354 = arith.constant 0 : i32
%355 = arith.extsi %354 : i32 to i64
%356 = llvm.getelementptr %arg6[%355] : (!llvm.ptr, i64) -> !llvm.ptr, i64
llvm.store %353, %356 : i64, !llvm.ptr
%357 = arith.constant 0 : i32
%358 = arith.extsi %357 : i32 to i64
%359 = llvm.getelementptr %arg7[%358] : (!llvm.ptr, i64) -> !llvm.ptr, i64
llvm.store %arg2, %359 : i64, !llvm.ptr
%360 = arith.constant 0 : i32
%361 = arith.extsi %360 : i32 to i64
%362 = llvm.getelementptr %arg8[%361] : (!llvm.ptr, i64) -> !llvm.ptr, i64
llvm.store %arg3, %362 : i64, !llvm.ptr
func.return
}
func.func @diamond_to_rect(%arg0: i64, %arg1: i64, %arg2: i64, %arg3: i64, %arg4: i64, %arg5: !llvm.ptr, %arg6: !llvm.ptr, %arg7: !llvm.ptr, %arg8: !llvm.ptr) -> () {
%364 = arith.subi %arg4, %arg3 : i64
%365 = arith.constant 2 : i32
%366 = arith.extsi %365 : i32 to i64
%363 = func.call @ceil_div(%364, %366) : (i64, i64) -> i64
%367 = arith.constant 0 : i32
%368 = arith.extsi %367 : i32 to i64
%369 = llvm.getelementptr %arg5[%368] : (!llvm.ptr, i64) -> !llvm.ptr, i64
llvm.store %363, %369 : i64, !llvm.ptr
%371 = arith.subi %arg4, %arg2 : i64
%372 = arith.constant 2 : i32
%373 = arith.extsi %372 : i32 to i64
%370 = func.call @floordiv(%371, %373) : (i64, i64) -> i64
%374 = arith.constant 0 : i32
%375 = arith.extsi %374 : i32 to i64
%376 = llvm.getelementptr %arg6[%375] : (!llvm.ptr, i64) -> !llvm.ptr, i64
llvm.store %370, %376 : i64, !llvm.ptr
%378 = arith.subi %arg4, %arg1 : i64
%379 = arith.constant 2 : i32
%380 = arith.extsi %379 : i32 to i64
%377 = func.call @ceil_div(%378, %380) : (i64, i64) -> i64
%381 = arith.constant 0 : i32
%382 = arith.extsi %381 : i32 to i64
%383 = llvm.getelementptr %arg7[%382] : (!llvm.ptr, i64) -> !llvm.ptr, i64
llvm.store %377, %383 : i64, !llvm.ptr
%385 = arith.subi %arg4, %arg0 : i64
%386 = arith.constant 2 : i32
%387 = arith.extsi %386 : i32 to i64
%384 = func.call @floordiv(%385, %387) : (i64, i64) -> i64
%388 = arith.constant 0 : i32
%389 = arith.extsi %388 : i32 to i64
%390 = llvm.getelementptr %arg8[%389] : (!llvm.ptr, i64) -> !llvm.ptr, i64
llvm.store %384, %390 : i64, !llvm.ptr
func.return
}
func.func @sum_region(%arg0: i32, %arg1: i64, %arg2: i64, %arg3: i64, %arg4: i64, %arg5: !llvm.ptr, %arg6: !llvm.ptr, %arg7: !llvm.ptr, %arg8: !llvm.ptr) -> i64 {
%391 = arith.cmpi sgt, %arg1, %arg2 : i64
%392 = scf.if %391 -> (i1) {
%393 = arith.constant true
scf.yield %393 : i1
} else {
%394 = arith.cmpi sgt, %arg3, %arg4 : i64
scf.yield %394 : i1
}
cf.cond_br %392, ^bb51, ^bb52
^bb51:
%395 = arith.constant 0 : i32
%396 = arith.extsi %395 : i32 to i64
func.return %396 : i64
^bb52:
cf.br ^bb53
^bb53:
%397 = arith.constant 0 : i32
%398 = arith.cmpi eq, %arg0, %397 : i32
cf.cond_br %398, ^bb54, ^bb55
^bb54:
%399 = func.call @count_rect(%arg1, %arg2, %arg3, %arg4) : (i64, i64, i64, i64) -> i128
%400 = llvm.mlir.addressof @BRUTE : !llvm.ptr
%401 = llvm.load %400 : !llvm.ptr -> i64
%402 = arith.extsi %401 : i64 to i128
%404 = arith.trunci %399 : i128 to i64
%405 = arith.trunci %402 : i128 to i64
%403 = arith.cmpi sle, %404, %405 : i64
cf.cond_br %403, ^bb57, ^bb58
^bb57:
%406 = func.call @brute_rect(%arg1, %arg2, %arg3, %arg4) : (i64, i64, i64, i64) -> i64
func.return %406 : i64
^bb58:
cf.br ^bb59
^bb59:
cf.br ^bb56
^bb55:
%407 = func.call @count_diamond(%arg1, %arg2, %arg3, %arg4) : (i64, i64, i64, i64) -> i128
%408 = llvm.mlir.addressof @BRUTE : !llvm.ptr
%409 = llvm.load %408 : !llvm.ptr -> i64
%410 = arith.extsi %409 : i64 to i128
%412 = arith.trunci %407 : i128 to i64
%413 = arith.trunci %410 : i128 to i64
%411 = arith.cmpi sle, %412, %413 : i64
cf.cond_br %411, ^bb60, ^bb61
^bb60:
%414 = func.call @brute_diamond(%arg1, %arg2, %arg3, %arg4) : (i64, i64, i64, i64) -> i64
func.return %414 : i64
^bb61:
cf.br ^bb62
^bb62:
cf.br ^bb56
^bb56:
%416 = arith.extsi %arg0 : i32 to i64
%415 = func.call @hkey(%arg1, %arg2, %arg3, %arg4, %416) : (i64, i64, i64, i64, i64) -> i64
%417 = func.call @hslot(%415, %arg5, %arg7) : (i64, !llvm.ptr, !llvm.ptr) -> i64
%419 = llvm.getelementptr %arg7[%417] : (!llvm.ptr, i64) -> !llvm.ptr, i8
%418 = llvm.load %419 : !llvm.ptr -> i8
%420 = arith.constant 1 : i32
%422 = arith.extsi %418 : i8 to i32
%421 = arith.cmpi eq, %422, %420 : i32
%423 = scf.if %421 -> (i1) {
%425 = llvm.getelementptr %arg5[%417] : (!llvm.ptr, i64) -> !llvm.ptr, i64
%424 = llvm.load %425 : !llvm.ptr -> i64
%426 = arith.cmpi eq, %424, %415 : i64
scf.yield %426 : i1
} else {
%427 = arith.constant false
scf.yield %427 : i1
}
cf.cond_br %423, ^bb63, ^bb64
^bb63:
%429 = llvm.getelementptr %arg6[%417] : (!llvm.ptr, i64) -> !llvm.ptr, i64
%428 = llvm.load %429 : !llvm.ptr -> i64
func.return %428 : i64
^bb64:
cf.br ^bb65
^bb65:
%430 = arith.constant 0 : i32
%431 = arith.extsi %430 : i32 to i64
%432 = llvm.mlir.constant(1 : i64) : i64
%433 = llvm.alloca %432 x i64 : (i64) -> !llvm.ptr
llvm.store %431, %433 : i64, !llvm.ptr
%434 = arith.constant 0 : i32
%435 = arith.cmpi eq, %arg0, %434 : i32
cf.cond_br %435, ^bb66, ^bb67
^bb66:
%437 = arith.constant 0 : i32
# String concatenation: !llvm.ptr + i32
# String concatenation: !llvm.ptr + i32
# String concatenation: !llvm.ptr + i32
%441 = arith.extsi %437 : i32 to i64
func.call @rect_to_diamond(%arg1, %arg2, %arg3, %arg4, %441, %arg8, %438, %439, %440) : (i64, i64, i64, i64, i64, !llvm.ptr, !llvm.ptr, !llvm.ptr, !llvm.ptr) -> ()
%443 = arith.constant 1 : i32
%445 = arith.constant 0 : i32
%446 = arith.extsi %445 : i32 to i64
%447 = llvm.getelementptr %arg8[%446] : (!llvm.ptr, i64) -> !llvm.ptr, i64
%444 = llvm.load %447 : !llvm.ptr -> i64
%449 = arith.constant 1 : i32
%450 = arith.extsi %449 : i32 to i64
%451 = llvm.getelementptr %arg8[%450] : (!llvm.ptr, i64) -> !llvm.ptr, i64
%448 = llvm.load %451 : !llvm.ptr -> i64
%453 = arith.constant 2 : i32
%454 = arith.extsi %453 : i32 to i64
%455 = llvm.getelementptr %arg8[%454] : (!llvm.ptr, i64) -> !llvm.ptr, i64
%452 = llvm.load %455 : !llvm.ptr -> i64
%457 = arith.constant 3 : i32
%458 = arith.extsi %457 : i32 to i64
%459 = llvm.getelementptr %arg8[%458] : (!llvm.ptr, i64) -> !llvm.ptr, i64
%456 = llvm.load %459 : !llvm.ptr -> i64
%442 = func.call @sum_region(%443, %444, %448, %452, %456, %arg5, %arg6, %arg7, %arg8) : (i32, i64, i64, i64, i64, !llvm.ptr, !llvm.ptr, !llvm.ptr, !llvm.ptr) -> i64
llvm.store %442, %433 : i64, !llvm.ptr
%461 = arith.constant 1 : i32
# String concatenation: !llvm.ptr + i32
# String concatenation: !llvm.ptr + i32
# String concatenation: !llvm.ptr + i32
%465 = arith.extsi %461 : i32 to i64
func.call @rect_to_diamond(%arg1, %arg2, %arg3, %arg4, %465, %arg8, %462, %463, %464) : (i64, i64, i64, i64, i64, !llvm.ptr, !llvm.ptr, !llvm.ptr, !llvm.ptr) -> ()
%467 = arith.constant 0 : i32
%468 = arith.extsi %467 : i32 to i64
%469 = llvm.getelementptr %arg8[%468] : (!llvm.ptr, i64) -> !llvm.ptr, i64
%466 = llvm.load %469 : !llvm.ptr -> i64
%471 = arith.constant 1 : i32
%472 = arith.extsi %471 : i32 to i64
%473 = llvm.getelementptr %arg8[%472] : (!llvm.ptr, i64) -> !llvm.ptr, i64
%470 = llvm.load %473 : !llvm.ptr -> i64
%475 = arith.constant 2 : i32
%476 = arith.extsi %475 : i32 to i64
%477 = llvm.getelementptr %arg8[%476] : (!llvm.ptr, i64) -> !llvm.ptr, i64
%474 = llvm.load %477 : !llvm.ptr -> i64
%479 = arith.constant 3 : i32
%480 = arith.extsi %479 : i32 to i64
%481 = llvm.getelementptr %arg8[%480] : (!llvm.ptr, i64) -> !llvm.ptr, i64
%478 = llvm.load %481 : !llvm.ptr -> i64
%482 = llvm.load %433 : !llvm.ptr -> i64
%484 = arith.constant 1 : i32
%483 = func.call @sum_region(%484, %466, %470, %474, %478, %arg5, %arg6, %arg7, %arg8) : (i32, i64, i64, i64, i64, !llvm.ptr, !llvm.ptr, !llvm.ptr, !llvm.ptr) -> i64
%485 = arith.addi %482, %483 : i64
%486 = llvm.mlir.addressof @MOD : !llvm.ptr
%487 = llvm.load %486 : !llvm.ptr -> i64
%488 = arith.remsi %485, %487 : i64
llvm.store %488, %433 : i64, !llvm.ptr
%489 = llvm.load %433 : !llvm.ptr -> i64
%490 = func.call @count_diamond(%466, %470, %474, %478) : (i64, i64, i64, i64) -> i128
%491 = llvm.mlir.addressof @MOD : !llvm.ptr
%492 = llvm.load %491 : !llvm.ptr -> i64
%493 = arith.extsi %492 : i64 to i128
%495 = arith.trunci %490 : i128 to i64
%496 = arith.trunci %493 : i128 to i64
%494 = arith.remsi %495, %496 : i64
%497 = arith.addi %489, %494 : i64
%498 = llvm.mlir.addressof @MOD : !llvm.ptr
%499 = llvm.load %498 : !llvm.ptr -> i64
%500 = arith.remsi %497, %499 : i64
llvm.store %500, %433 : i64, !llvm.ptr
cf.br ^bb68
^bb67:
%502 = arith.constant 0 : i32
# String concatenation: !llvm.ptr + i32
# String concatenation: !llvm.ptr + i32
# String concatenation: !llvm.ptr + i32
%506 = arith.extsi %502 : i32 to i64
func.call @diamond_to_rect(%arg1, %arg2, %arg3, %arg4, %506, %arg8, %503, %504, %505) : (i64, i64, i64, i64, i64, !llvm.ptr, !llvm.ptr, !llvm.ptr, !llvm.ptr) -> ()
%508 = arith.constant 0 : i32
%510 = arith.constant 0 : i32
%511 = arith.extsi %510 : i32 to i64
%512 = llvm.getelementptr %arg8[%511] : (!llvm.ptr, i64) -> !llvm.ptr, i64
%509 = llvm.load %512 : !llvm.ptr -> i64
%514 = arith.constant 1 : i32
%515 = arith.extsi %514 : i32 to i64
%516 = llvm.getelementptr %arg8[%515] : (!llvm.ptr, i64) -> !llvm.ptr, i64
%513 = llvm.load %516 : !llvm.ptr -> i64
%518 = arith.constant 2 : i32
%519 = arith.extsi %518 : i32 to i64
%520 = llvm.getelementptr %arg8[%519] : (!llvm.ptr, i64) -> !llvm.ptr, i64
%517 = llvm.load %520 : !llvm.ptr -> i64
%522 = arith.constant 3 : i32
%523 = arith.extsi %522 : i32 to i64
%524 = llvm.getelementptr %arg8[%523] : (!llvm.ptr, i64) -> !llvm.ptr, i64
%521 = llvm.load %524 : !llvm.ptr -> i64
%507 = func.call @sum_region(%508, %509, %513, %517, %521, %arg5, %arg6, %arg7, %arg8) : (i32, i64, i64, i64, i64, !llvm.ptr, !llvm.ptr, !llvm.ptr, !llvm.ptr) -> i64
llvm.store %507, %433 : i64, !llvm.ptr
%526 = arith.constant 1 : i32
# String concatenation: !llvm.ptr + i32
# String concatenation: !llvm.ptr + i32
# String concatenation: !llvm.ptr + i32
%530 = arith.extsi %526 : i32 to i64
func.call @diamond_to_rect(%arg1, %arg2, %arg3, %arg4, %530, %arg8, %527, %528, %529) : (i64, i64, i64, i64, i64, !llvm.ptr, !llvm.ptr, !llvm.ptr, !llvm.ptr) -> ()
%532 = arith.constant 0 : i32
%533 = arith.extsi %532 : i32 to i64
%534 = llvm.getelementptr %arg8[%533] : (!llvm.ptr, i64) -> !llvm.ptr, i64
%531 = llvm.load %534 : !llvm.ptr -> i64
%536 = arith.constant 1 : i32
%537 = arith.extsi %536 : i32 to i64
%538 = llvm.getelementptr %arg8[%537] : (!llvm.ptr, i64) -> !llvm.ptr, i64
%535 = llvm.load %538 : !llvm.ptr -> i64
%540 = arith.constant 2 : i32
%541 = arith.extsi %540 : i32 to i64
%542 = llvm.getelementptr %arg8[%541] : (!llvm.ptr, i64) -> !llvm.ptr, i64
%539 = llvm.load %542 : !llvm.ptr -> i64
%544 = arith.constant 3 : i32
%545 = arith.extsi %544 : i32 to i64
%546 = llvm.getelementptr %arg8[%545] : (!llvm.ptr, i64) -> !llvm.ptr, i64
%543 = llvm.load %546 : !llvm.ptr -> i64
%547 = llvm.load %433 : !llvm.ptr -> i64
%549 = arith.constant 0 : i32
%548 = func.call @sum_region(%549, %531, %535, %539, %543, %arg5, %arg6, %arg7, %arg8) : (i32, i64, i64, i64, i64, !llvm.ptr, !llvm.ptr, !llvm.ptr, !llvm.ptr) -> i64
%550 = arith.addi %547, %548 : i64
%551 = llvm.mlir.addressof @MOD : !llvm.ptr
%552 = llvm.load %551 : !llvm.ptr -> i64
%553 = arith.remsi %550, %552 : i64
llvm.store %553, %433 : i64, !llvm.ptr
%554 = llvm.load %433 : !llvm.ptr -> i64
%555 = func.call @count_rect(%531, %535, %539, %543) : (i64, i64, i64, i64) -> i128
%556 = llvm.mlir.addressof @MOD : !llvm.ptr
%557 = llvm.load %556 : !llvm.ptr -> i64
%558 = arith.extsi %557 : i64 to i128
%560 = arith.trunci %555 : i128 to i64
%561 = arith.trunci %558 : i128 to i64
%559 = arith.remsi %560, %561 : i64
%562 = arith.addi %554, %559 : i64
%563 = llvm.mlir.addressof @MOD : !llvm.ptr
%564 = llvm.load %563 : !llvm.ptr -> i64
%565 = arith.remsi %562, %564 : i64
llvm.store %565, %433 : i64, !llvm.ptr
cf.br ^bb68
^bb68:
%566 = func.call @hslot(%415, %arg5, %arg7) : (i64, !llvm.ptr, !llvm.ptr) -> i64
%567 = arith.constant 1 : i32
%568 = arith.trunci %567 : i32 to i8
%569 = llvm.getelementptr %arg7[%566] : (!llvm.ptr, i64) -> !llvm.ptr, i8
llvm.store %568, %569 : i8, !llvm.ptr
%570 = llvm.getelementptr %arg5[%566] : (!llvm.ptr, i64) -> !llvm.ptr, i64
llvm.store %415, %570 : i64, !llvm.ptr
%571 = llvm.load %433 : !llvm.ptr -> i64
%572 = llvm.getelementptr %arg6[%566] : (!llvm.ptr, i64) -> !llvm.ptr, i64
llvm.store %571, %572 : i64, !llvm.ptr
%573 = llvm.load %433 : !llvm.ptr -> i64
func.return %573 : i64
}
func.func @main() -> i32 {
%575 = llvm.mlir.addressof @CAP : !llvm.ptr
%576 = llvm.load %575 : !llvm.ptr -> i64
%577 = arith.constant 8 : i32
%578 = arith.extsi %577 : i32 to i64
%574 = func.call @calloc(%576, %578) : (i64, i64) -> !llvm.ptr
%580 = llvm.mlir.addressof @CAP : !llvm.ptr
%581 = llvm.load %580 : !llvm.ptr -> i64
%582 = arith.constant 8 : i32
%583 = arith.extsi %582 : i32 to i64
%579 = func.call @calloc(%581, %583) : (i64, i64) -> !llvm.ptr
%585 = llvm.mlir.addressof @CAP : !llvm.ptr
%586 = llvm.load %585 : !llvm.ptr -> i64
%587 = arith.constant 1 : i32
%588 = arith.extsi %587 : i32 to i64
%584 = func.call @calloc(%586, %588) : (i64, i64) -> !llvm.ptr
%590 = arith.constant 4 : i32
%591 = arith.constant 8 : i32
%592 = arith.extsi %590 : i32 to i64
%593 = arith.extsi %591 : i32 to i64
%589 = func.call @calloc(%592, %593) : (i64, i64) -> !llvm.ptr
%594 = llvm.mlir.zero : !llvm.ptr
%595 = llvm.icmp "eq" %574, %594 : !llvm.ptr
%596 = scf.if %595 -> (i1) {
%597 = arith.constant true
scf.yield %597 : i1
} else {
%598 = llvm.mlir.zero : !llvm.ptr
%599 = llvm.icmp "eq" %579, %598 : !llvm.ptr
scf.yield %599 : i1
}
%600 = scf.if %596 -> (i1) {
%601 = arith.constant true
scf.yield %601 : i1
} else {
%602 = llvm.mlir.zero : !llvm.ptr
%603 = llvm.icmp "eq" %584, %602 : !llvm.ptr
scf.yield %603 : i1
}
%604 = scf.if %600 -> (i1) {
%605 = arith.constant true
scf.yield %605 : i1
} else {
%606 = llvm.mlir.zero : !llvm.ptr
%607 = llvm.icmp "eq" %589, %606 : !llvm.ptr
scf.yield %607 : i1
}
cf.cond_br %604, ^bb69, ^bb70
^bb69:
%608 = arith.constant 1 : i32
func.return %608 : i32
^bb70:
cf.br ^bb71
^bb71:
%610 = arith.constant 0 : i32
%611 = llvm.mlir.addressof @LIM : !llvm.ptr
%612 = llvm.load %611 : !llvm.ptr -> i64
%614 = arith.constant 0 : i64
%613 = arith.subi %614, %612 : i64
%615 = llvm.mlir.addressof @LIM : !llvm.ptr
%616 = llvm.load %615 : !llvm.ptr -> i64
%617 = llvm.mlir.addressof @LIM : !llvm.ptr
%618 = llvm.load %617 : !llvm.ptr -> i64
%620 = arith.constant 0 : i64
%619 = arith.subi %620, %618 : i64
%621 = llvm.mlir.addressof @LIM : !llvm.ptr
%622 = llvm.load %621 : !llvm.ptr -> i64
%609 = func.call @sum_region(%610, %613, %616, %619, %622, %574, %579, %584, %589) : (i32, i64, i64, i64, i64, !llvm.ptr, !llvm.ptr, !llvm.ptr, !llvm.ptr) -> i64
%623 = llvm.mlir.addressof @str_0 : !llvm.ptr
%624 = llvm.call @printf(%623, %609) vararg(!llvm.func<i32 (ptr, ...)>) : (!llvm.ptr, i64) -> i32
func.call @free(%574) : (!llvm.ptr) -> ()
func.call @free(%579) : (!llvm.ptr) -> ()
func.call @free(%584) : (!llvm.ptr) -> ()
func.call @free(%589) : (!llvm.ptr) -> ()
%629 = arith.constant 0 : i32
func.return %629 : i32
}
}