Problem 675

2^omega(n): sum of S(i!) for i=2..10^7, where S(n) = prod_{p^e || n} (2e+1), mod 1e9+87.

Answer416146418
Output416146418
StatusPASS
Native helperno
Runtime2270 ms
Peak memory318784 KB
Time complexityO(n^3) (estimated)
Space complexityO(n^2) (estimated)

Performance comparison

MetricOur solutionBest known
Time complexityO(n^3)O(n log log n)
Space complexityO(n^2)O(n)
ApproachFlow solutionSieve or enumeration
VerdictSuboptimal

Flow source

# Project Euler 675
# 2^omega(n): sum of S(i!) for i=2..10^7, where S(n) = prod_{p^e || n} (2e+1), mod 1e9+87.

extern {
    function calloc(n: i64, size: i64) -> ptr<void>
    function free(p: ptr<void>) -> void
}

const MOD: i64 = 1000000087

function mod_pow(base0: i64, exp0: i64, mod: i64) -> i64 {
    let mut r: i64 = 1
    let mut b: i64 = base0 % mod
    let mut e: i64 = exp0
    while e > 0 {
        if (e & 1) == 1 {
            let t: i128 = (r as i128) * (b as i128) % (mod as i128)
            r = t as i64
        }
        let t2: i128 = (b as i128) * (b as i128) % (mod as i128)
        b = t2 as i64
        e = e / 2
    }
    return r
}

function main() -> i32 {
    let N: i64 = 10000000

    # Smallest prime factor sieve
    let spf: ptr<i64> = calloc(N + 1, 8)
    let primes: ptr<i64> = calloc(N / 5 + 100, 8)
    let mut npc: i64 = 0

    if spf == null || primes == null { return 1 }

    for x in 2..(N + 1) {
        if spf[x] == 0 {
            spf[x] = x
            primes[npc] = x
            npc = npc + 1
        }
        let spfx: i64 = spf[x]
        for j in 0..npc {
            let p: i64 = primes[j]
            let y: i64 = p * x
            if y > N { break }
            spf[y] = p
            if p == spfx { break }
        }
    }

    # Compute v_2(N!) to size the inverse table
    let mut v2_fact: i64 = 0
    let mut power: i64 = 2
    while power <= N {
        v2_fact = v2_fact + N / power
        power = power * 2
    }

    # Inverse table up to 2*v2_fact + 1
    let inv_limit: i64 = 2 * v2_fact + 1
    let inv: ptr<i64> = calloc(inv_limit + 1, 8)
    if inv == null { return 1 }
    inv[1] = 1
    for x in 2..(inv_limit + 1) {
        inv[x] = (MOD - (MOD / x) * inv[MOD % x] % MOD) % MOD
    }

    # Exponents of each prime in the running factorial
    let exps: ptr<i64> = calloc(N + 1, 8)
    if exps == null { return 1 }

    let mut running: i64 = 1
    let mut total: i64 = 0

    for i in 2..(N + 1) {
        let mut x: i64 = i
        while x > 1 {
            let p: i64 = spf[x]
            let mut count: i64 = 0
            while x % p == 0 {
                x = x / p
                count = count + 1
            }

            let old_exp: i64 = exps[p]
            let new_exp: i64 = old_exp + count
            let old_factor: i64 = 2 * old_exp + 1
            let new_factor: i64 = 2 * new_exp + 1
            let t1: i128 = (running as i128) * (new_factor as i128) % (MOD as i128)
            let t2: i128 = (t1 * (inv[old_factor] as i128)) % (MOD as i128)
            running = t2 as i64
            exps[p] = new_exp
        }

        total = (total + running) % MOD
    }

    printf("%lld\n", total)

    free(exps)
    free(inv)
    free(primes)
    free(spf)
    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 mod_pow_i64_i64_i64(int64_t base0, int64_t exp0, int64_t mod);
int32_t main(void);

static const int64_t MOD = 1000000087;



int64_t mod_pow_i64_i64_i64(int64_t base0, int64_t exp0, int64_t mod) {
    int64_t r = 1;
    int64_t b = FLOW_CHECKED_MOD((base0), (mod));
    int64_t e = exp0;
    while (e > 0) {
        if ((e & 1) == 1) {
            __int128 t = FLOW_CHECKED_MOD(((((__int128)(r)) * ((__int128)(b)))), (((__int128)(mod))));
            r = ((int64_t)(t));
        }
        __int128 t2 = FLOW_CHECKED_MOD(((((__int128)(b)) * ((__int128)(b)))), (((__int128)(mod))));
        b = ((int64_t)(t2));
        e = FLOW_CHECKED_DIV((e), (2));
    }
    return r;
}

int32_t main(void) {
    int64_t N = 10000000;
    int64_t* spf = (int64_t*)(calloc((N + 1), 8));
    int64_t* primes = (int64_t*)(calloc((FLOW_CHECKED_DIV((N), (5)) + 100), 8));
    int64_t npc = 0;
    if ((spf == NULL || primes == NULL)) {
        return 1;
    }
    int32_t __flow_step_1 = 1;
    for (int32_t x = 2; (2 <= (N + 1)) ? x < (N + 1) : x > (N + 1); x += (2 <= (N + 1)) ? 1 : -1) {
        if (spf[x] == 0) {
            spf[x] = x;
            primes[npc] = x;
            npc = (npc + 1);
        }
        int64_t spfx = spf[x];
        int32_t __flow_step_2 = 1;
        for (int32_t j = 0; (0 <= npc) ? j < npc : j > npc; j += (0 <= npc) ? 1 : -1) {
            int64_t p = primes[j];
            int64_t y = (p * x);
            if (y > N) {
                break;
            }
            spf[y] = p;
            if (p == spfx) {
                break;
            }
        }
    }
    int64_t v2_fact = 0;
    int64_t power = 2;
    while (power <= N) {
        v2_fact = (v2_fact + FLOW_CHECKED_DIV((N), (power)));
        power = (power * 2);
    }
    int64_t inv_limit = ((2 * v2_fact) + 1);
    int64_t* inv = (int64_t*)(calloc((inv_limit + 1), 8));
    if (inv == NULL) {
        return 1;
    }
    inv[1] = 1;
    int32_t __flow_step_3 = 1;
    for (int32_t x = 2; (2 <= (inv_limit + 1)) ? x < (inv_limit + 1) : x > (inv_limit + 1); x += (2 <= (inv_limit + 1)) ? 1 : -1) {
        inv[x] = FLOW_CHECKED_MOD(((MOD - FLOW_CHECKED_MOD(((FLOW_CHECKED_DIV((MOD), (x)) * inv[FLOW_CHECKED_MOD((MOD), (x))])), (MOD)))), (MOD));
    }
    int64_t* exps = (int64_t*)(calloc((N + 1), 8));
    if (exps == NULL) {
        return 1;
    }
    int64_t running = 1;
    int64_t total = 0;
    int32_t __flow_step_4 = 1;
    for (int32_t i = 2; (2 <= (N + 1)) ? i < (N + 1) : i > (N + 1); i += (2 <= (N + 1)) ? 1 : -1) {
        int64_t x = i;
        while (x > 1) {
            int64_t p = spf[x];
            int64_t count = 0;
            while (FLOW_CHECKED_MOD((x), (p)) == 0) {
                x = FLOW_CHECKED_DIV((x), (p));
                count = (count + 1);
            }
            int64_t old_exp = exps[p];
            int64_t new_exp = (old_exp + count);
            int64_t old_factor = ((2 * old_exp) + 1);
            int64_t new_factor = ((2 * new_exp) + 1);
            __int128 t1 = FLOW_CHECKED_MOD(((((__int128)(running)) * ((__int128)(new_factor)))), (((__int128)(MOD))));
            __int128 t2 = FLOW_CHECKED_MOD(((t1 * ((__int128)(inv[old_factor])))), (((__int128)(MOD))));
            running = ((int64_t)(t2));
            exps[p] = new_exp;
        }
        total = FLOW_CHECKED_MOD(((total + running)), (MOD));
    }
    printf("%lld\n", total);
    free(exps);
    free(inv);
    free(primes);
    free(spf);
    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>
  func.func private @calloc(i64, i64) -> !llvm.ptr
  func.func private @free(!llvm.ptr) -> ()
  // Constant: MOD
  llvm.mlir.global internal constant @MOD(1000000087 : i64) : i64
  func.func @mod_pow(%arg0: i64, %arg1: i64, %arg2: i64) -> i64 {
    %0 = arith.constant 1 : i32
    %1 = arith.extsi %0 : i32 to i64
    %2 = llvm.mlir.constant(1 : i64) : i64
    %3 = llvm.alloca %2 x i64 : (i64) -> !llvm.ptr
    llvm.store %1, %3 : i64, !llvm.ptr
    %4 = arith.remsi %arg0, %arg2 : i64
    %5 = llvm.mlir.constant(1 : i64) : i64
    %6 = llvm.alloca %5 x i64 : (i64) -> !llvm.ptr
    llvm.store %4, %6 : i64, !llvm.ptr
    %7 = llvm.mlir.constant(1 : i64) : i64
    %8 = llvm.alloca %7 x i64 : (i64) -> !llvm.ptr
    llvm.store %arg1, %8 : i64, !llvm.ptr
    cf.br ^bb0
    ^bb0:
    %9 = llvm.load %8 : !llvm.ptr -> i64
    %10 = arith.constant 0 : i32
    %12 = arith.extsi %10 : i32 to i64
    %11 = arith.cmpi sgt, %9, %12 : i64
    cf.cond_br %11, ^bb1, ^bb2
    ^bb1:
      %13 = llvm.load %8 : !llvm.ptr -> i64
      %14 = arith.constant 1 : i32
      %16 = arith.extsi %14 : i32 to i64
      %15 = arith.andi %13, %16 : i64
      %17 = arith.constant 1 : i32
      %19 = arith.extsi %17 : i32 to i64
      %18 = arith.cmpi eq, %15, %19 : i64
      cf.cond_br %18, ^bb3, ^bb4
      ^bb3:
        %20 = llvm.load %3 : !llvm.ptr -> i64
        %21 = arith.extsi %20 : i64 to i128
        %22 = llvm.load %6 : !llvm.ptr -> i64
        %23 = arith.extsi %22 : i64 to i128
        %25 = arith.trunci %21 : i128 to i64
        %26 = arith.trunci %23 : i128 to i64
        %24 = arith.muli %25, %26 : i64
        %27 = arith.extsi %arg2 : i64 to i128
        %29 = arith.trunci %27 : i128 to i64
        %28 = arith.remsi %24, %29 : i64
        %30 = arith.extsi %28 : i64 to i128
        %31 = arith.trunci %30 : i128 to i64
        llvm.store %31, %3 : i64, !llvm.ptr
        cf.br ^bb5
      ^bb4:
        cf.br ^bb5
      ^bb5:
      %32 = llvm.load %6 : !llvm.ptr -> i64
      %33 = arith.extsi %32 : i64 to i128
      %34 = llvm.load %6 : !llvm.ptr -> i64
      %35 = arith.extsi %34 : i64 to i128
      %37 = arith.trunci %33 : i128 to i64
      %38 = arith.trunci %35 : i128 to i64
      %36 = arith.muli %37, %38 : i64
      %39 = arith.extsi %arg2 : i64 to i128
      %41 = arith.trunci %39 : i128 to i64
      %40 = arith.remsi %36, %41 : i64
      %42 = arith.extsi %40 : i64 to i128
      %43 = arith.trunci %42 : i128 to i64
      llvm.store %43, %6 : i64, !llvm.ptr
      %44 = llvm.load %8 : !llvm.ptr -> i64
      %45 = arith.constant 2 : i32
      %47 = arith.extsi %45 : i32 to i64
      %46 = arith.divsi %44, %47 : i64
      llvm.store %46, %8 : i64, !llvm.ptr
      cf.br ^bb0
    ^bb2:
    %48 = llvm.load %3 : !llvm.ptr -> i64
    func.return %48 : i64
  }
  func.func @main() -> i32 {
    %49 = arith.constant 10000000 : i32
    %50 = arith.extsi %49 : i32 to i64
    %52 = arith.constant 1 : i32
    %54 = arith.extsi %52 : i32 to i64
    %53 = arith.addi %50, %54 : i64
    %55 = arith.constant 8 : i32
    %56 = arith.extsi %55 : i32 to i64
    %51 = func.call @calloc(%53, %56) : (i64, i64) -> !llvm.ptr
    %58 = arith.constant 5 : i32
    %60 = arith.extsi %58 : i32 to i64
    %59 = arith.divsi %50, %60 : i64
    %61 = arith.constant 100 : i32
    %63 = arith.extsi %61 : i32 to i64
    %62 = arith.addi %59, %63 : i64
    %64 = arith.constant 8 : i32
    %65 = arith.extsi %64 : i32 to i64
    %57 = func.call @calloc(%62, %65) : (i64, i64) -> !llvm.ptr
    %66 = arith.constant 0 : i32
    %67 = arith.extsi %66 : i32 to i64
    %68 = llvm.mlir.constant(1 : i64) : i64
    %69 = llvm.alloca %68 x i64 : (i64) -> !llvm.ptr
    llvm.store %67, %69 : i64, !llvm.ptr
    %70 = llvm.mlir.zero : !llvm.ptr
    %71 = llvm.icmp "eq" %51, %70 : !llvm.ptr
    %72 = scf.if %71 -> (i1) {
      %73 = arith.constant true
      scf.yield %73 : i1
    } else {
      %74 = llvm.mlir.zero : !llvm.ptr
      %75 = llvm.icmp "eq" %57, %74 : !llvm.ptr
      scf.yield %75 : i1
    }
    cf.cond_br %72, ^bb6, ^bb7
    ^bb6:
      %76 = arith.constant 1 : i32
      func.return %76 : i32
    ^bb7:
      cf.br ^bb8
    ^bb8:
    %77 = arith.constant 2 : i32
    %78 = arith.constant 1 : i32
    %80 = arith.extsi %78 : i32 to i64
    %79 = arith.addi %50, %80 : i64
    %81 = arith.index_cast %77 : i32 to index
    %82 = arith.index_cast %79 : i32 to index
    %84 = arith.constant 1 : index
    %85 = arith.constant -1 : index
    %86 = arith.cmpi sle, %81, %82 : index
    %83 = arith.select %86, %84, %85 : index
    cf.br ^bb9(%81 : index)
    ^bb9(%87: index):
    %88 = arith.cmpi slt, %87, %82 : index
    %89 = arith.cmpi sgt, %87, %82 : index
    %90 = arith.select %86, %88, %89 : i1
    cf.cond_br %90, ^bb10(%87 : index), ^bb11(%87 : index)
    ^bb10(%91: index):
      %93 = arith.index_cast %91 : index to i64
      %94 = llvm.getelementptr %51[%93] : (!llvm.ptr, i64) -> !llvm.ptr, i64
      %92 = llvm.load %94 : !llvm.ptr -> i64
      %95 = arith.constant 0 : i32
      %97 = arith.extsi %95 : i32 to i64
      %96 = arith.cmpi eq, %92, %97 : i64
      cf.cond_br %96, ^bb12, ^bb13
      ^bb12:
        %98 = arith.index_cast %91 : index to i64
        %99 = arith.index_cast %91 : index to i64
        %100 = llvm.getelementptr %51[%99] : (!llvm.ptr, i64) -> !llvm.ptr, i64
        llvm.store %98, %100 : i64, !llvm.ptr
        %101 = llvm.load %69 : !llvm.ptr -> i64
        %102 = arith.index_cast %91 : index to i64
        %103 = llvm.getelementptr %57[%101] : (!llvm.ptr, i64) -> !llvm.ptr, i64
        llvm.store %102, %103 : i64, !llvm.ptr
        %104 = llvm.load %69 : !llvm.ptr -> i64
        %105 = arith.constant 1 : i32
        %107 = arith.extsi %105 : i32 to i64
        %106 = arith.addi %104, %107 : i64
        llvm.store %106, %69 : i64, !llvm.ptr
        cf.br ^bb14
      ^bb13:
        cf.br ^bb14
      ^bb14:
      %109 = arith.index_cast %91 : index to i64
      %110 = llvm.getelementptr %51[%109] : (!llvm.ptr, i64) -> !llvm.ptr, i64
      %108 = llvm.load %110 : !llvm.ptr -> i64
      %111 = arith.constant 0 : i32
      %112 = llvm.load %69 : !llvm.ptr -> i64
      %113 = arith.index_cast %111 : i32 to index
      %114 = arith.index_cast %112 : i32 to index
      %116 = arith.constant 1 : index
      %117 = arith.constant -1 : index
      %118 = arith.cmpi sle, %113, %114 : index
      %115 = arith.select %118, %116, %117 : index
      cf.br ^bb15(%113 : index)
      ^bb15(%119: index):
      %120 = arith.cmpi slt, %119, %114 : index
      %121 = arith.cmpi sgt, %119, %114 : index
      %122 = arith.select %118, %120, %121 : i1
      cf.cond_br %122, ^bb16(%119 : index), ^bb17(%119 : index)
      ^bb16(%123: index):
        %125 = arith.index_cast %123 : index to i64
        %126 = llvm.getelementptr %57[%125] : (!llvm.ptr, i64) -> !llvm.ptr, i64
        %124 = llvm.load %126 : !llvm.ptr -> i64
        %128 = arith.trunci %124 : i64 to i32
        %129 = arith.index_cast %91 : index to i32
        %127 = arith.muli %128, %129 : i32
        %130 = arith.extsi %127 : i32 to i64
        %131 = arith.cmpi sgt, %130, %50 : i64
        cf.cond_br %131, ^bb18, ^bb19
        ^bb18:
          cf.br ^bb17(%123 : index)
        ^bb19:
          cf.br ^bb20
        ^bb20:
        %132 = llvm.getelementptr %51[%130] : (!llvm.ptr, i64) -> !llvm.ptr, i64
        llvm.store %124, %132 : i64, !llvm.ptr
        %133 = arith.cmpi eq, %124, %108 : i64
        cf.cond_br %133, ^bb21, ^bb22
        ^bb21:
          cf.br ^bb17(%123 : index)
        ^bb22:
          cf.br ^bb23
        ^bb23:
        %134 = arith.addi %123, %115 : index
        cf.br ^bb15(%134 : index)
      ^bb17(%135: index):
      %136 = arith.addi %91, %83 : index
      cf.br ^bb9(%136 : index)
    ^bb11(%137: index):
    %138 = arith.constant 0 : i32
    %139 = arith.extsi %138 : i32 to i64
    %140 = llvm.mlir.constant(1 : i64) : i64
    %141 = llvm.alloca %140 x i64 : (i64) -> !llvm.ptr
    llvm.store %139, %141 : i64, !llvm.ptr
    %142 = arith.constant 2 : i32
    %143 = arith.extsi %142 : i32 to i64
    %144 = llvm.mlir.constant(1 : i64) : i64
    %145 = llvm.alloca %144 x i64 : (i64) -> !llvm.ptr
    llvm.store %143, %145 : i64, !llvm.ptr
    cf.br ^bb24
    ^bb24:
    %146 = llvm.load %145 : !llvm.ptr -> i64
    %147 = arith.cmpi sle, %146, %50 : i64
    cf.cond_br %147, ^bb25, ^bb26
    ^bb25:
      %148 = llvm.load %141 : !llvm.ptr -> i64
      %149 = llvm.load %145 : !llvm.ptr -> i64
      %150 = arith.divsi %50, %149 : i64
      %151 = arith.addi %148, %150 : i64
      llvm.store %151, %141 : i64, !llvm.ptr
      %152 = llvm.load %145 : !llvm.ptr -> i64
      %153 = arith.constant 2 : i32
      %155 = arith.extsi %153 : i32 to i64
      %154 = arith.muli %152, %155 : i64
      llvm.store %154, %145 : i64, !llvm.ptr
      cf.br ^bb24
    ^bb26:
    %156 = arith.constant 2 : i32
    %157 = llvm.load %141 : !llvm.ptr -> i64
    %159 = arith.extsi %156 : i32 to i64
    %158 = arith.muli %159, %157 : i64
    %160 = arith.constant 1 : i32
    %162 = arith.extsi %160 : i32 to i64
    %161 = arith.addi %158, %162 : i64
    %164 = arith.constant 1 : i32
    %166 = arith.extsi %164 : i32 to i64
    %165 = arith.addi %161, %166 : i64
    %167 = arith.constant 8 : i32
    %168 = arith.extsi %167 : i32 to i64
    %163 = func.call @calloc(%165, %168) : (i64, i64) -> !llvm.ptr
    %169 = llvm.mlir.zero : !llvm.ptr
    %170 = llvm.icmp "eq" %163, %169 : !llvm.ptr
    cf.cond_br %170, ^bb27, ^bb28
    ^bb27:
      %171 = arith.constant 1 : i32
      func.return %171 : i32
    ^bb28:
      cf.br ^bb29
    ^bb29:
    %172 = arith.constant 1 : i32
    %173 = arith.constant 1 : i32
    %174 = arith.extsi %172 : i32 to i64
    %175 = arith.extsi %173 : i32 to i64
    %176 = llvm.getelementptr %163[%175] : (!llvm.ptr, i64) -> !llvm.ptr, i64
    llvm.store %174, %176 : i64, !llvm.ptr
    %177 = arith.constant 2 : i32
    %178 = arith.constant 1 : i32
    %180 = arith.extsi %178 : i32 to i64
    %179 = arith.addi %161, %180 : i64
    %181 = arith.index_cast %177 : i32 to index
    %182 = arith.index_cast %179 : i32 to index
    %184 = arith.constant 1 : index
    %185 = arith.constant -1 : index
    %186 = arith.cmpi sle, %181, %182 : index
    %183 = arith.select %186, %184, %185 : index
    cf.br ^bb30(%181 : index)
    ^bb30(%187: index):
    %188 = arith.cmpi slt, %187, %182 : index
    %189 = arith.cmpi sgt, %187, %182 : index
    %190 = arith.select %186, %188, %189 : i1
    cf.cond_br %190, ^bb31(%187 : index), ^bb32(%187 : index)
    ^bb31(%191: index):
      %192 = llvm.mlir.addressof @MOD : !llvm.ptr
      %193 = llvm.load %192 : !llvm.ptr -> i64
      %194 = llvm.mlir.addressof @MOD : !llvm.ptr
      %195 = llvm.load %194 : !llvm.ptr -> i64
      %197 = arith.trunci %195 : i64 to i32
      %198 = arith.index_cast %191 : index to i32
      %196 = arith.divsi %197, %198 : i32
      %200 = llvm.mlir.addressof @MOD : !llvm.ptr
      %201 = llvm.load %200 : !llvm.ptr -> i64
      %203 = arith.trunci %201 : i64 to i32
      %204 = arith.index_cast %191 : index to i32
      %202 = arith.remsi %203, %204 : i32
      %205 = arith.extsi %202 : i32 to i64
      %206 = llvm.getelementptr %163[%205] : (!llvm.ptr, i64) -> !llvm.ptr, i64
      %199 = llvm.load %206 : !llvm.ptr -> i64
      %208 = arith.extsi %196 : i32 to i64
      %207 = arith.muli %208, %199 : i64
      %209 = llvm.mlir.addressof @MOD : !llvm.ptr
      %210 = llvm.load %209 : !llvm.ptr -> i64
      %211 = arith.remsi %207, %210 : i64
      %212 = arith.subi %193, %211 : i64
      %213 = llvm.mlir.addressof @MOD : !llvm.ptr
      %214 = llvm.load %213 : !llvm.ptr -> i64
      %215 = arith.remsi %212, %214 : i64
      %216 = arith.index_cast %191 : index to i64
      %217 = llvm.getelementptr %163[%216] : (!llvm.ptr, i64) -> !llvm.ptr, i64
      llvm.store %215, %217 : i64, !llvm.ptr
      %218 = arith.addi %191, %183 : index
      cf.br ^bb30(%218 : index)
    ^bb32(%219: index):
    %221 = arith.constant 1 : i32
    %223 = arith.extsi %221 : i32 to i64
    %222 = arith.addi %50, %223 : i64
    %224 = arith.constant 8 : i32
    %225 = arith.extsi %224 : i32 to i64
    %220 = func.call @calloc(%222, %225) : (i64, i64) -> !llvm.ptr
    %226 = llvm.mlir.zero : !llvm.ptr
    %227 = llvm.icmp "eq" %220, %226 : !llvm.ptr
    cf.cond_br %227, ^bb33, ^bb34
    ^bb33:
      %228 = arith.constant 1 : i32
      func.return %228 : i32
    ^bb34:
      cf.br ^bb35
    ^bb35:
    %229 = arith.constant 1 : i32
    %230 = arith.extsi %229 : i32 to 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 = arith.constant 0 : i32
    %234 = arith.extsi %233 : i32 to i64
    %235 = llvm.mlir.constant(1 : i64) : i64
    %236 = llvm.alloca %235 x i64 : (i64) -> !llvm.ptr
    llvm.store %234, %236 : i64, !llvm.ptr
    %237 = arith.constant 2 : i32
    %238 = arith.constant 1 : i32
    %240 = arith.extsi %238 : i32 to i64
    %239 = arith.addi %50, %240 : i64
    %241 = arith.index_cast %237 : i32 to index
    %242 = arith.index_cast %239 : i32 to index
    %244 = arith.constant 1 : index
    %245 = arith.constant -1 : index
    %246 = arith.cmpi sle, %241, %242 : index
    %243 = arith.select %246, %244, %245 : index
    cf.br ^bb36(%241 : index)
    ^bb36(%247: index):
    %248 = arith.cmpi slt, %247, %242 : index
    %249 = arith.cmpi sgt, %247, %242 : index
    %250 = arith.select %246, %248, %249 : i1
    cf.cond_br %250, ^bb37(%247 : index), ^bb38(%247 : index)
    ^bb37(%251: index):
      %252 = arith.index_cast %251 : index to i64
      %253 = llvm.mlir.constant(1 : i64) : i64
      %254 = llvm.alloca %253 x i64 : (i64) -> !llvm.ptr
      llvm.store %252, %254 : i64, !llvm.ptr
      cf.br ^bb39
      ^bb39:
      %255 = llvm.load %254 : !llvm.ptr -> i64
      %256 = arith.constant 1 : i32
      %258 = arith.extsi %256 : i32 to i64
      %257 = arith.cmpi sgt, %255, %258 : i64
      cf.cond_br %257, ^bb40, ^bb41
      ^bb40:
        %260 = llvm.load %254 : !llvm.ptr -> i64
        %261 = llvm.getelementptr %51[%260] : (!llvm.ptr, i64) -> !llvm.ptr, i64
        %259 = llvm.load %261 : !llvm.ptr -> i64
        %262 = arith.constant 0 : i32
        %263 = arith.extsi %262 : i32 to i64
        %264 = llvm.mlir.constant(1 : i64) : i64
        %265 = llvm.alloca %264 x i64 : (i64) -> !llvm.ptr
        llvm.store %263, %265 : i64, !llvm.ptr
        cf.br ^bb42
        ^bb42:
        %266 = llvm.load %254 : !llvm.ptr -> i64
        %267 = arith.remsi %266, %259 : i64
        %268 = arith.constant 0 : i32
        %270 = arith.extsi %268 : i32 to i64
        %269 = arith.cmpi eq, %267, %270 : i64
        cf.cond_br %269, ^bb43, ^bb44
        ^bb43:
          %271 = llvm.load %254 : !llvm.ptr -> i64
          %272 = arith.divsi %271, %259 : i64
          llvm.store %272, %254 : i64, !llvm.ptr
          %273 = llvm.load %265 : !llvm.ptr -> i64
          %274 = arith.constant 1 : i32
          %276 = arith.extsi %274 : i32 to i64
          %275 = arith.addi %273, %276 : i64
          llvm.store %275, %265 : i64, !llvm.ptr
          cf.br ^bb42
        ^bb44:
        %278 = llvm.getelementptr %220[%259] : (!llvm.ptr, i64) -> !llvm.ptr, i64
        %277 = llvm.load %278 : !llvm.ptr -> i64
        %279 = llvm.load %265 : !llvm.ptr -> i64
        %280 = arith.addi %277, %279 : i64
        %281 = arith.constant 2 : i32
        %283 = arith.extsi %281 : i32 to i64
        %282 = arith.muli %283, %277 : i64
        %284 = arith.constant 1 : i32
        %286 = arith.extsi %284 : i32 to i64
        %285 = arith.addi %282, %286 : i64
        %287 = arith.constant 2 : i32
        %289 = arith.extsi %287 : i32 to i64
        %288 = arith.muli %289, %280 : i64
        %290 = arith.constant 1 : i32
        %292 = arith.extsi %290 : i32 to i64
        %291 = arith.addi %288, %292 : i64
        %293 = llvm.load %232 : !llvm.ptr -> i64
        %294 = arith.extsi %293 : i64 to i128
        %295 = arith.extsi %291 : i64 to i128
        %297 = arith.trunci %294 : i128 to i64
        %298 = arith.trunci %295 : i128 to i64
        %296 = arith.muli %297, %298 : i64
        %299 = llvm.mlir.addressof @MOD : !llvm.ptr
        %300 = llvm.load %299 : !llvm.ptr -> i64
        %301 = arith.extsi %300 : i64 to i128
        %303 = arith.trunci %301 : i128 to i64
        %302 = arith.remsi %296, %303 : i64
        %304 = arith.extsi %302 : i64 to i128
        %306 = llvm.getelementptr %163[%285] : (!llvm.ptr, i64) -> !llvm.ptr, i64
        %305 = llvm.load %306 : !llvm.ptr -> i64
        %307 = arith.extsi %305 : i64 to i128
        %309 = arith.trunci %304 : i128 to i64
        %310 = arith.trunci %307 : i128 to i64
        %308 = arith.muli %309, %310 : i64
        %311 = llvm.mlir.addressof @MOD : !llvm.ptr
        %312 = llvm.load %311 : !llvm.ptr -> i64
        %313 = arith.extsi %312 : i64 to i128
        %315 = arith.trunci %313 : i128 to i64
        %314 = arith.remsi %308, %315 : i64
        %316 = arith.extsi %314 : i64 to i128
        %317 = arith.trunci %316 : i128 to i64
        llvm.store %317, %232 : i64, !llvm.ptr
        %318 = llvm.getelementptr %220[%259] : (!llvm.ptr, i64) -> !llvm.ptr, i64
        llvm.store %280, %318 : i64, !llvm.ptr
        cf.br ^bb39
      ^bb41:
      %319 = llvm.load %236 : !llvm.ptr -> i64
      %320 = llvm.load %232 : !llvm.ptr -> i64
      %321 = arith.addi %319, %320 : i64
      %322 = llvm.mlir.addressof @MOD : !llvm.ptr
      %323 = llvm.load %322 : !llvm.ptr -> i64
      %324 = arith.remsi %321, %323 : i64
      llvm.store %324, %236 : i64, !llvm.ptr
      %325 = arith.addi %251, %243 : index
      cf.br ^bb36(%325 : index)
    ^bb38(%326: index):
    %327 = llvm.mlir.addressof @str_0 : !llvm.ptr
    %328 = llvm.load %236 : !llvm.ptr -> i64
    %329 = llvm.call @printf(%327, %328) vararg(!llvm.func<i32 (ptr, ...)>) : (!llvm.ptr, i64) -> i32
    func.call @free(%220) : (!llvm.ptr) -> ()
    func.call @free(%163) : (!llvm.ptr) -> ()
    func.call @free(%57) : (!llvm.ptr) -> ()
    func.call @free(%51) : (!llvm.ptr) -> ()
    %334 = arith.constant 0 : i32
    func.return %334 : i32
  }
}