Problem 171

Last 9 digits of sum of all n < 10^20 whose digit-square-sum is a square.

Answer142989277
Output142989277
StatusPASS
Native helperno
Runtime50 ms
Peak memory1664 KB
Time complexityO(n^3) (estimated)
Space complexityO(n) (estimated)

Performance comparison

MetricOur solutionBest known
Time complexityO(n^3)O(n)
Space complexityO(n)O(n)
ApproachFlow solutionBig-integer arithmetic
VerdictSuboptimal

Flow source

# Project Euler 171
# Last 9 digits of sum of all n < 10^20 whose digit-square-sum is a square.

import euler.nt { isqrt, mod_pow }

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

function main() -> i32 {
    let length: i32 = 20
    let mod: i64 = 1000000000
    let max_s: i32 = 81 * length  # 9^2 * 20

    let sqsum: ptr<i64> = calloc((length + 1) as i64 * (max_s + 1) as i64, 8)
    let count: ptr<i64> = calloc((length + 1) as i64 * (max_s + 1) as i64, 8)
    if sqsum == null || count == null { return 1 }

    # index: i*(max_s+1) + s
    count[0] = 1

    let mut i: i32 = 1
    while i <= length {
        let mut j: i32 = 0
        while j < 10 {
            let j2: i32 = j * j
            let mut k: i32 = 0
            while k + j2 <= max_s {
                let idx: i64 = (i as i64) * ((max_s + 1) as i64) + ((k + j2) as i64)
                let prev: i64 = ((i - 1) as i64) * ((max_s + 1) as i64) + (k as i64)
                let add: i64 = (sqsum[prev] + mod_pow(10, (i - 1) as i64, mod) * (j as i64) % mod * count[prev] % mod) % mod
                sqsum[idx] = (sqsum[idx] + add) % mod
                count[idx] = (count[idx] + count[prev]) % mod
                k = k + 1
            }
            j = j + 1
        }
        i = i + 1
    }

    let mut ans: i64 = 0
    let max_root: i64 = isqrt(max_s as i64)
    let mut r: i64 = 1
    while r <= max_root {
        let s: i64 = r * r
        let idx: i64 = (length as i64) * ((max_s + 1) as i64) + s
        ans = (ans + sqsum[idx]) % mod
        r = r + 1
    }
    printf("%lld\n", ans)
    free(sqsum)
    free(count)
    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 gcd_i64_i64(int64_t a0, int64_t b0);
int64_t lcm_i64_i64(int64_t a, int64_t b);
int64_t isqrt_i64(int64_t n);
int64_t mulmod_i64_i64_i64(int64_t a0, int64_t b0, int64_t mod);
int64_t mod_pow_i64_i64_i64(int64_t base, int64_t exp, int64_t mod);
bool is_prime_i64(int64_t n);
int32_t main(void);

int64_t gcd_i64_i64(int64_t a0, int64_t b0) {
    int64_t a = a0;
    int64_t b = b0;
    while (b != 0) {
        int64_t t = FLOW_CHECKED_MOD((a), (b));
        a = b;
        b = t;
    }
    return a;
}

int64_t lcm_i64_i64(int64_t a, int64_t b) {
    if ((a == 0 || b == 0)) {
        return 0;
    }
    return (FLOW_CHECKED_DIV((a), (gcd_i64_i64(a, b))) * b);
}

int64_t isqrt_i64(int64_t n) {
    if (n < 2) {
        return n;
    }
    int64_t x = n;
    int64_t y = FLOW_CHECKED_DIV(((x + 1)), (2));
    while (y < x) {
        x = y;
        y = FLOW_CHECKED_DIV(((x + FLOW_CHECKED_DIV((n), (x)))), (2));
    }
    return x;
}

int64_t mulmod_i64_i64_i64(int64_t a0, int64_t b0, int64_t mod) {
    int64_t a = FLOW_CHECKED_MOD((a0), (mod));
    int64_t b = FLOW_CHECKED_MOD((b0), (mod));
    int64_t result = 0;
    while (b > 0) {
        if (FLOW_CHECKED_MOD((b), (2)) == 1) {
            result = FLOW_CHECKED_MOD(((result + a)), (mod));
        }
        a = FLOW_CHECKED_MOD(((a * 2)), (mod));
        b = FLOW_CHECKED_DIV((b), (2));
    }
    return result;
}

int64_t mod_pow_i64_i64_i64(int64_t base, int64_t exp, int64_t mod) {
    if (mod == 1) {
        return 0;
    }
    int64_t result = 1;
    int64_t b = FLOW_CHECKED_MOD((base), (mod));
    int64_t e = exp;
    while (e > 0) {
        if (FLOW_CHECKED_MOD((e), (2)) == 1) {
            result = mulmod_i64_i64_i64(result, b, mod);
        }
        b = mulmod_i64_i64_i64(b, b, mod);
        e = FLOW_CHECKED_DIV((e), (2));
    }
    return result;
}

bool is_prime_i64(int64_t n) {
    if (n < 2) {
        return 0;
    }
    if (n < 4) {
        return 1;
    }
    if ((FLOW_CHECKED_MOD((n), (2)) == 0 || FLOW_CHECKED_MOD((n), (3)) == 0)) {
        return 0;
    }
    int64_t i = 5;
    while ((i * i) <= n) {
        if ((FLOW_CHECKED_MOD((n), (i)) == 0 || FLOW_CHECKED_MOD((n), ((i + 2))) == 0)) {
            return 0;
        }
        i = (i + 6);
    }
    return 1;
}



int32_t main(void) {
    int32_t length = 20;
    int64_t mod = 1000000000;
    int32_t max_s = (81 * length);
    int64_t* sqsum = (int64_t*)(calloc((((int64_t)((length + 1))) * ((int64_t)((max_s + 1)))), 8));
    int64_t* count = (int64_t*)(calloc((((int64_t)((length + 1))) * ((int64_t)((max_s + 1)))), 8));
    if ((sqsum == NULL || count == NULL)) {
        return 1;
    }
    count[0] = 1;
    int32_t i = 1;
    while (i <= length) {
        int32_t j = 0;
        while (j < 10) {
            int32_t j2 = (j * j);
            int32_t k = 0;
            while ((k + j2) <= max_s) {
                int64_t idx = ((((int64_t)(i)) * ((int64_t)((max_s + 1)))) + ((int64_t)((k + j2))));
                int64_t prev = ((((int64_t)((i - 1))) * ((int64_t)((max_s + 1)))) + ((int64_t)(k)));
                int64_t add = FLOW_CHECKED_MOD(((sqsum[prev] + FLOW_CHECKED_MOD(((FLOW_CHECKED_MOD(((mod_pow_i64_i64_i64(10, ((int64_t)((i - 1))), mod) * ((int64_t)(j)))), (mod)) * count[prev])), (mod)))), (mod));
                sqsum[idx] = FLOW_CHECKED_MOD(((sqsum[idx] + add)), (mod));
                count[idx] = FLOW_CHECKED_MOD(((count[idx] + count[prev])), (mod));
                k = (k + 1);
            }
            j = (j + 1);
        }
        i = (i + 1);
    }
    int64_t ans = 0;
    int64_t max_root = isqrt_i64(((int64_t)(max_s)));
    int64_t r = 1;
    while (r <= max_root) {
        int64_t s = (r * r);
        int64_t idx = ((((int64_t)(length)) * ((int64_t)((max_s + 1)))) + s);
        ans = FLOW_CHECKED_MOD(((ans + sqsum[idx])), (mod));
        r = (r + 1);
    }
    printf("%lld\n", ans);
    free(sqsum);
    free(count);
    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 @gcd(%arg0: i64, %arg1: i64) -> i64 {
    %0 = llvm.mlir.constant(1 : i64) : i64
    %1 = llvm.alloca %0 x i64 : (i64) -> !llvm.ptr
    llvm.store %arg0, %1 : i64, !llvm.ptr
    %2 = llvm.mlir.constant(1 : i64) : i64
    %3 = llvm.alloca %2 x i64 : (i64) -> !llvm.ptr
    llvm.store %arg1, %3 : i64, !llvm.ptr
    cf.br ^bb0
    ^bb0:
    %4 = llvm.load %3 : !llvm.ptr -> i64
    %5 = arith.constant 0 : i32
    %7 = arith.extsi %5 : i32 to i64
    %6 = arith.cmpi ne, %4, %7 : i64
    cf.cond_br %6, ^bb1, ^bb2
    ^bb1:
      %8 = llvm.load %1 : !llvm.ptr -> i64
      %9 = llvm.load %3 : !llvm.ptr -> i64
      %10 = arith.remsi %8, %9 : i64
      %11 = llvm.load %3 : !llvm.ptr -> i64
      llvm.store %11, %1 : i64, !llvm.ptr
      llvm.store %10, %3 : i64, !llvm.ptr
      cf.br ^bb0
    ^bb2:
    %12 = llvm.load %1 : !llvm.ptr -> i64
    func.return %12 : i64
  }
  func.func @lcm(%arg0: i64, %arg1: i64) -> i64 {
    %13 = arith.constant 0 : i32
    %15 = arith.extsi %13 : i32 to i64
    %14 = arith.cmpi eq, %arg0, %15 : i64
    %16 = scf.if %14 -> (i1) {
      %17 = arith.constant true
      scf.yield %17 : i1
    } else {
      %18 = arith.constant 0 : i32
      %20 = arith.extsi %18 : i32 to i64
      %19 = arith.cmpi eq, %arg1, %20 : i64
      scf.yield %19 : i1
    }
    cf.cond_br %16, ^bb3, ^bb4
    ^bb3:
      %21 = arith.constant 0 : i32
      %22 = arith.extsi %21 : i32 to i64
      func.return %22 : i64
    ^bb4:
      cf.br ^bb5
    ^bb5:
    %23 = func.call @gcd(%arg0, %arg1) : (i64, i64) -> i64
    %24 = arith.divsi %arg0, %23 : i64
    %25 = arith.muli %24, %arg1 : i64
    func.return %25 : i64
  }
  func.func @isqrt(%arg0: i64) -> i64 {
    %26 = arith.constant 2 : i32
    %28 = arith.extsi %26 : i32 to i64
    %27 = arith.cmpi slt, %arg0, %28 : i64
    cf.cond_br %27, ^bb6, ^bb7
    ^bb6:
      func.return %arg0 : i64
    ^bb7:
      cf.br ^bb8
    ^bb8:
    %29 = llvm.mlir.constant(1 : i64) : i64
    %30 = llvm.alloca %29 x i64 : (i64) -> !llvm.ptr
    llvm.store %arg0, %30 : i64, !llvm.ptr
    %31 = llvm.load %30 : !llvm.ptr -> i64
    %32 = arith.constant 1 : i32
    %34 = arith.extsi %32 : i32 to i64
    %33 = arith.addi %31, %34 : i64
    %35 = arith.constant 2 : i32
    %37 = arith.extsi %35 : i32 to i64
    %36 = arith.divsi %33, %37 : i64
    %38 = llvm.mlir.constant(1 : i64) : i64
    %39 = llvm.alloca %38 x i64 : (i64) -> !llvm.ptr
    llvm.store %36, %39 : i64, !llvm.ptr
    cf.br ^bb9
    ^bb9:
    %40 = llvm.load %39 : !llvm.ptr -> i64
    %41 = llvm.load %30 : !llvm.ptr -> i64
    %42 = arith.cmpi slt, %40, %41 : i64
    cf.cond_br %42, ^bb10, ^bb11
    ^bb10:
      %43 = llvm.load %39 : !llvm.ptr -> i64
      llvm.store %43, %30 : i64, !llvm.ptr
      %44 = llvm.load %30 : !llvm.ptr -> i64
      %45 = llvm.load %30 : !llvm.ptr -> i64
      %46 = arith.divsi %arg0, %45 : i64
      %47 = arith.addi %44, %46 : i64
      %48 = arith.constant 2 : i32
      %50 = arith.extsi %48 : i32 to i64
      %49 = arith.divsi %47, %50 : i64
      llvm.store %49, %39 : i64, !llvm.ptr
      cf.br ^bb9
    ^bb11:
    %51 = llvm.load %30 : !llvm.ptr -> i64
    func.return %51 : i64
  }
  func.func @mulmod(%arg0: i64, %arg1: i64, %arg2: i64) -> i64 {
    %52 = arith.remsi %arg0, %arg2 : i64
    %53 = llvm.mlir.constant(1 : i64) : i64
    %54 = llvm.alloca %53 x i64 : (i64) -> !llvm.ptr
    llvm.store %52, %54 : i64, !llvm.ptr
    %55 = arith.remsi %arg1, %arg2 : i64
    %56 = llvm.mlir.constant(1 : i64) : i64
    %57 = llvm.alloca %56 x i64 : (i64) -> !llvm.ptr
    llvm.store %55, %57 : i64, !llvm.ptr
    %58 = arith.constant 0 : i32
    %59 = arith.extsi %58 : i32 to i64
    %60 = llvm.mlir.constant(1 : i64) : i64
    %61 = llvm.alloca %60 x i64 : (i64) -> !llvm.ptr
    llvm.store %59, %61 : i64, !llvm.ptr
    cf.br ^bb12
    ^bb12:
    %62 = llvm.load %57 : !llvm.ptr -> i64
    %63 = arith.constant 0 : i32
    %65 = arith.extsi %63 : i32 to i64
    %64 = arith.cmpi sgt, %62, %65 : i64
    cf.cond_br %64, ^bb13, ^bb14
    ^bb13:
      %66 = llvm.load %57 : !llvm.ptr -> i64
      %67 = arith.constant 2 : i32
      %69 = arith.extsi %67 : i32 to i64
      %68 = arith.remsi %66, %69 : i64
      %70 = arith.constant 1 : i32
      %72 = arith.extsi %70 : i32 to i64
      %71 = arith.cmpi eq, %68, %72 : i64
      cf.cond_br %71, ^bb15, ^bb16
      ^bb15:
        %73 = llvm.load %61 : !llvm.ptr -> i64
        %74 = llvm.load %54 : !llvm.ptr -> i64
        %75 = arith.addi %73, %74 : i64
        %76 = arith.remsi %75, %arg2 : i64
        llvm.store %76, %61 : i64, !llvm.ptr
        cf.br ^bb17
      ^bb16:
        cf.br ^bb17
      ^bb17:
      %77 = llvm.load %54 : !llvm.ptr -> i64
      %78 = arith.constant 2 : i32
      %80 = arith.extsi %78 : i32 to i64
      %79 = arith.muli %77, %80 : i64
      %81 = arith.remsi %79, %arg2 : i64
      llvm.store %81, %54 : i64, !llvm.ptr
      %82 = llvm.load %57 : !llvm.ptr -> i64
      %83 = arith.constant 2 : i32
      %85 = arith.extsi %83 : i32 to i64
      %84 = arith.divsi %82, %85 : i64
      llvm.store %84, %57 : i64, !llvm.ptr
      cf.br ^bb12
    ^bb14:
    %86 = llvm.load %61 : !llvm.ptr -> i64
    func.return %86 : i64
  }
  func.func @mod_pow(%arg0: i64, %arg1: i64, %arg2: i64) -> i64 {
    %87 = arith.constant 1 : i32
    %89 = arith.extsi %87 : i32 to i64
    %88 = arith.cmpi eq, %arg2, %89 : i64
    cf.cond_br %88, ^bb18, ^bb19
    ^bb18:
      %90 = arith.constant 0 : i32
      %91 = arith.extsi %90 : i32 to i64
      func.return %91 : i64
    ^bb19:
      cf.br ^bb20
    ^bb20:
    %92 = arith.constant 1 : i32
    %93 = arith.extsi %92 : i32 to i64
    %94 = llvm.mlir.constant(1 : i64) : i64
    %95 = llvm.alloca %94 x i64 : (i64) -> !llvm.ptr
    llvm.store %93, %95 : i64, !llvm.ptr
    %96 = arith.remsi %arg0, %arg2 : i64
    %97 = llvm.mlir.constant(1 : i64) : i64
    %98 = llvm.alloca %97 x i64 : (i64) -> !llvm.ptr
    llvm.store %96, %98 : i64, !llvm.ptr
    %99 = llvm.mlir.constant(1 : i64) : i64
    %100 = llvm.alloca %99 x i64 : (i64) -> !llvm.ptr
    llvm.store %arg1, %100 : i64, !llvm.ptr
    cf.br ^bb21
    ^bb21:
    %101 = llvm.load %100 : !llvm.ptr -> i64
    %102 = arith.constant 0 : i32
    %104 = arith.extsi %102 : i32 to i64
    %103 = arith.cmpi sgt, %101, %104 : i64
    cf.cond_br %103, ^bb22, ^bb23
    ^bb22:
      %105 = llvm.load %100 : !llvm.ptr -> i64
      %106 = arith.constant 2 : i32
      %108 = arith.extsi %106 : i32 to i64
      %107 = arith.remsi %105, %108 : i64
      %109 = arith.constant 1 : i32
      %111 = arith.extsi %109 : i32 to i64
      %110 = arith.cmpi eq, %107, %111 : i64
      cf.cond_br %110, ^bb24, ^bb25
      ^bb24:
        %113 = llvm.load %95 : !llvm.ptr -> i64
        %114 = llvm.load %98 : !llvm.ptr -> i64
        %112 = func.call @mulmod(%113, %114, %arg2) : (i64, i64, i64) -> i64
        llvm.store %112, %95 : i64, !llvm.ptr
        cf.br ^bb26
      ^bb25:
        cf.br ^bb26
      ^bb26:
      %116 = llvm.load %98 : !llvm.ptr -> i64
      %117 = llvm.load %98 : !llvm.ptr -> i64
      %115 = func.call @mulmod(%116, %117, %arg2) : (i64, i64, i64) -> i64
      llvm.store %115, %98 : i64, !llvm.ptr
      %118 = llvm.load %100 : !llvm.ptr -> i64
      %119 = arith.constant 2 : i32
      %121 = arith.extsi %119 : i32 to i64
      %120 = arith.divsi %118, %121 : i64
      llvm.store %120, %100 : i64, !llvm.ptr
      cf.br ^bb21
    ^bb23:
    %122 = llvm.load %95 : !llvm.ptr -> i64
    func.return %122 : i64
  }
  func.func @is_prime(%arg0: i64) -> i1 {
    %123 = arith.constant 2 : i32
    %125 = arith.extsi %123 : i32 to i64
    %124 = arith.cmpi slt, %arg0, %125 : i64
    cf.cond_br %124, ^bb27, ^bb28
    ^bb27:
      %126 = arith.constant 0 : i1
      func.return %126 : i1
    ^bb28:
      cf.br ^bb29
    ^bb29:
    %127 = arith.constant 4 : i32
    %129 = arith.extsi %127 : i32 to i64
    %128 = arith.cmpi slt, %arg0, %129 : i64
    cf.cond_br %128, ^bb30, ^bb31
    ^bb30:
      %130 = arith.constant 1 : i1
      func.return %130 : i1
    ^bb31:
      cf.br ^bb32
    ^bb32:
    %131 = arith.constant 2 : i32
    %133 = arith.extsi %131 : i32 to i64
    %132 = arith.remsi %arg0, %133 : i64
    %134 = arith.constant 0 : i32
    %136 = arith.extsi %134 : i32 to i64
    %135 = arith.cmpi eq, %132, %136 : i64
    %137 = scf.if %135 -> (i1) {
      %138 = arith.constant true
      scf.yield %138 : i1
    } else {
      %139 = arith.constant 3 : i32
      %141 = arith.extsi %139 : i32 to i64
      %140 = arith.remsi %arg0, %141 : i64
      %142 = arith.constant 0 : i32
      %144 = arith.extsi %142 : i32 to i64
      %143 = arith.cmpi eq, %140, %144 : i64
      scf.yield %143 : i1
    }
    cf.cond_br %137, ^bb33, ^bb34
    ^bb33:
      %145 = arith.constant 0 : i1
      func.return %145 : i1
    ^bb34:
      cf.br ^bb35
    ^bb35:
    %146 = arith.constant 5 : i32
    %147 = arith.extsi %146 : i32 to i64
    %148 = llvm.mlir.constant(1 : i64) : i64
    %149 = llvm.alloca %148 x i64 : (i64) -> !llvm.ptr
    llvm.store %147, %149 : i64, !llvm.ptr
    cf.br ^bb36
    ^bb36:
    %150 = llvm.load %149 : !llvm.ptr -> i64
    %151 = llvm.load %149 : !llvm.ptr -> i64
    %152 = arith.muli %150, %151 : i64
    %153 = arith.cmpi sle, %152, %arg0 : i64
    cf.cond_br %153, ^bb37, ^bb38
    ^bb37:
      %154 = llvm.load %149 : !llvm.ptr -> i64
      %155 = arith.remsi %arg0, %154 : i64
      %156 = arith.constant 0 : i32
      %158 = arith.extsi %156 : i32 to i64
      %157 = arith.cmpi eq, %155, %158 : i64
      %159 = scf.if %157 -> (i1) {
        %160 = arith.constant true
        scf.yield %160 : i1
      } else {
        %161 = llvm.load %149 : !llvm.ptr -> i64
        %162 = arith.constant 2 : i32
        %164 = arith.extsi %162 : i32 to i64
        %163 = arith.addi %161, %164 : i64
        %165 = arith.remsi %arg0, %163 : i64
        %166 = arith.constant 0 : i32
        %168 = arith.extsi %166 : i32 to i64
        %167 = arith.cmpi eq, %165, %168 : i64
        scf.yield %167 : i1
      }
      cf.cond_br %159, ^bb39, ^bb40
      ^bb39:
        %169 = arith.constant 0 : i1
        func.return %169 : i1
      ^bb40:
        cf.br ^bb41
      ^bb41:
      %170 = llvm.load %149 : !llvm.ptr -> i64
      %171 = arith.constant 6 : i32
      %173 = arith.extsi %171 : i32 to i64
      %172 = arith.addi %170, %173 : i64
      llvm.store %172, %149 : i64, !llvm.ptr
      cf.br ^bb36
    ^bb38:
    %174 = arith.constant 1 : i1
    func.return %174 : i1
  }
  func.func private @calloc(i64, i64) -> !llvm.ptr
  func.func private @free(!llvm.ptr) -> ()
  func.func @main() -> i32 {
    %175 = arith.constant 20 : i32
    %176 = arith.constant 1000000000 : i32
    %177 = arith.extsi %176 : i32 to i64
    %178 = arith.constant 81 : i32
    %179 = arith.muli %178, %175 : i32
    %181 = arith.constant 1 : i32
    %182 = arith.addi %175, %181 : i32
    %183 = arith.extsi %182 : i32 to i64
    %184 = arith.constant 1 : i32
    %185 = arith.addi %179, %184 : i32
    %186 = arith.extsi %185 : i32 to i64
    %187 = arith.muli %183, %186 : i64
    %188 = arith.constant 8 : i32
    %189 = arith.extsi %188 : i32 to i64
    %180 = func.call @calloc(%187, %189) : (i64, i64) -> !llvm.ptr
    %191 = arith.constant 1 : i32
    %192 = arith.addi %175, %191 : i32
    %193 = arith.extsi %192 : i32 to i64
    %194 = arith.constant 1 : i32
    %195 = arith.addi %179, %194 : i32
    %196 = arith.extsi %195 : i32 to i64
    %197 = arith.muli %193, %196 : i64
    %198 = arith.constant 8 : i32
    %199 = arith.extsi %198 : i32 to i64
    %190 = func.call @calloc(%197, %199) : (i64, i64) -> !llvm.ptr
    %200 = llvm.mlir.zero : !llvm.ptr
    %201 = llvm.icmp "eq" %180, %200 : !llvm.ptr
    %202 = scf.if %201 -> (i1) {
      %203 = arith.constant true
      scf.yield %203 : i1
    } else {
      %204 = llvm.mlir.zero : !llvm.ptr
      %205 = llvm.icmp "eq" %190, %204 : !llvm.ptr
      scf.yield %205 : i1
    }
    cf.cond_br %202, ^bb42, ^bb43
    ^bb42:
      %206 = arith.constant 1 : i32
      func.return %206 : i32
    ^bb43:
      cf.br ^bb44
    ^bb44:
    %207 = arith.constant 1 : i32
    %208 = arith.constant 0 : i32
    %209 = arith.extsi %207 : i32 to i64
    %210 = arith.extsi %208 : i32 to i64
    %211 = llvm.getelementptr %190[%210] : (!llvm.ptr, i64) -> !llvm.ptr, i64
    llvm.store %209, %211 : i64, !llvm.ptr
    %212 = arith.constant 1 : i32
    %213 = llvm.mlir.constant(1 : i64) : i64
    %214 = llvm.alloca %213 x i32 : (i64) -> !llvm.ptr
    llvm.store %212, %214 : i32, !llvm.ptr
    cf.br ^bb45
    ^bb45:
    %215 = llvm.load %214 : !llvm.ptr -> i32
    %216 = arith.cmpi sle, %215, %175 : i32
    cf.cond_br %216, ^bb46, ^bb47
    ^bb46:
      %217 = arith.constant 0 : i32
      %218 = llvm.mlir.constant(1 : i64) : i64
      %219 = llvm.alloca %218 x i32 : (i64) -> !llvm.ptr
      llvm.store %217, %219 : i32, !llvm.ptr
      cf.br ^bb48
      ^bb48:
      %220 = llvm.load %219 : !llvm.ptr -> i32
      %221 = arith.constant 10 : i32
      %222 = arith.cmpi slt, %220, %221 : i32
      cf.cond_br %222, ^bb49, ^bb50
      ^bb49:
        %223 = llvm.load %219 : !llvm.ptr -> i32
        %224 = llvm.load %219 : !llvm.ptr -> i32
        %225 = arith.muli %223, %224 : i32
        %226 = arith.constant 0 : i32
        %227 = llvm.mlir.constant(1 : i64) : i64
        %228 = llvm.alloca %227 x i32 : (i64) -> !llvm.ptr
        llvm.store %226, %228 : i32, !llvm.ptr
        cf.br ^bb51
        ^bb51:
        %229 = llvm.load %228 : !llvm.ptr -> i32
        %230 = arith.addi %229, %225 : i32
        %231 = arith.cmpi sle, %230, %179 : i32
        cf.cond_br %231, ^bb52, ^bb53
        ^bb52:
          %232 = llvm.load %214 : !llvm.ptr -> i32
          %233 = arith.extsi %232 : i32 to i64
          %234 = arith.constant 1 : i32
          %235 = arith.addi %179, %234 : i32
          %236 = arith.extsi %235 : i32 to i64
          %237 = arith.muli %233, %236 : i64
          %238 = llvm.load %228 : !llvm.ptr -> i32
          %239 = arith.addi %238, %225 : i32
          %240 = arith.extsi %239 : i32 to i64
          %241 = arith.addi %237, %240 : i64
          %242 = llvm.load %214 : !llvm.ptr -> i32
          %243 = arith.constant 1 : i32
          %244 = arith.subi %242, %243 : i32
          %245 = arith.extsi %244 : i32 to i64
          %246 = arith.constant 1 : i32
          %247 = arith.addi %179, %246 : i32
          %248 = arith.extsi %247 : i32 to i64
          %249 = arith.muli %245, %248 : i64
          %250 = llvm.load %228 : !llvm.ptr -> i32
          %251 = arith.extsi %250 : i32 to i64
          %252 = arith.addi %249, %251 : i64
          %254 = llvm.getelementptr %180[%252] : (!llvm.ptr, i64) -> !llvm.ptr, i64
          %253 = llvm.load %254 : !llvm.ptr -> i64
          %256 = arith.constant 10 : i32
          %257 = llvm.load %214 : !llvm.ptr -> i32
          %258 = arith.constant 1 : i32
          %259 = arith.subi %257, %258 : i32
          %260 = arith.extsi %259 : i32 to i64
          %261 = arith.extsi %256 : i32 to i64
          %255 = func.call @mod_pow(%261, %260, %177) : (i64, i64, i64) -> i64
          %262 = llvm.load %219 : !llvm.ptr -> i32
          %263 = arith.extsi %262 : i32 to i64
          %264 = arith.muli %255, %263 : i64
          %265 = arith.remsi %264, %177 : i64
          %267 = llvm.getelementptr %190[%252] : (!llvm.ptr, i64) -> !llvm.ptr, i64
          %266 = llvm.load %267 : !llvm.ptr -> i64
          %268 = arith.muli %265, %266 : i64
          %269 = arith.remsi %268, %177 : i64
          %270 = arith.addi %253, %269 : i64
          %271 = arith.remsi %270, %177 : i64
          %273 = llvm.getelementptr %180[%241] : (!llvm.ptr, i64) -> !llvm.ptr, i64
          %272 = llvm.load %273 : !llvm.ptr -> i64
          %274 = arith.addi %272, %271 : i64
          %275 = arith.remsi %274, %177 : i64
          %276 = llvm.getelementptr %180[%241] : (!llvm.ptr, i64) -> !llvm.ptr, i64
          llvm.store %275, %276 : i64, !llvm.ptr
          %278 = llvm.getelementptr %190[%241] : (!llvm.ptr, i64) -> !llvm.ptr, i64
          %277 = llvm.load %278 : !llvm.ptr -> i64
          %280 = llvm.getelementptr %190[%252] : (!llvm.ptr, i64) -> !llvm.ptr, i64
          %279 = llvm.load %280 : !llvm.ptr -> i64
          %281 = arith.addi %277, %279 : i64
          %282 = arith.remsi %281, %177 : i64
          %283 = llvm.getelementptr %190[%241] : (!llvm.ptr, i64) -> !llvm.ptr, i64
          llvm.store %282, %283 : i64, !llvm.ptr
          %284 = llvm.load %228 : !llvm.ptr -> i32
          %285 = arith.constant 1 : i32
          %286 = arith.addi %284, %285 : i32
          llvm.store %286, %228 : i32, !llvm.ptr
          cf.br ^bb51
        ^bb53:
        %287 = llvm.load %219 : !llvm.ptr -> i32
        %288 = arith.constant 1 : i32
        %289 = arith.addi %287, %288 : i32
        llvm.store %289, %219 : i32, !llvm.ptr
        cf.br ^bb48
      ^bb50:
      %290 = llvm.load %214 : !llvm.ptr -> i32
      %291 = arith.constant 1 : i32
      %292 = arith.addi %290, %291 : i32
      llvm.store %292, %214 : i32, !llvm.ptr
      cf.br ^bb45
    ^bb47:
    %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
    %298 = arith.extsi %179 : i32 to i64
    %297 = func.call @isqrt(%298) : (i64) -> i64
    %299 = arith.constant 1 : i32
    %300 = arith.extsi %299 : i32 to i64
    %301 = llvm.mlir.constant(1 : i64) : i64
    %302 = llvm.alloca %301 x i64 : (i64) -> !llvm.ptr
    llvm.store %300, %302 : i64, !llvm.ptr
    cf.br ^bb54
    ^bb54:
    %303 = llvm.load %302 : !llvm.ptr -> i64
    %304 = arith.cmpi sle, %303, %297 : i64
    cf.cond_br %304, ^bb55, ^bb56
    ^bb55:
      %305 = llvm.load %302 : !llvm.ptr -> i64
      %306 = llvm.load %302 : !llvm.ptr -> i64
      %307 = arith.muli %305, %306 : i64
      %308 = arith.extsi %175 : i32 to i64
      %309 = arith.constant 1 : i32
      %310 = arith.addi %179, %309 : i32
      %311 = arith.extsi %310 : i32 to i64
      %312 = arith.muli %308, %311 : i64
      %313 = arith.addi %312, %307 : i64
      %314 = llvm.load %296 : !llvm.ptr -> i64
      %316 = llvm.getelementptr %180[%313] : (!llvm.ptr, i64) -> !llvm.ptr, i64
      %315 = llvm.load %316 : !llvm.ptr -> i64
      %317 = arith.addi %314, %315 : i64
      %318 = arith.remsi %317, %177 : i64
      llvm.store %318, %296 : i64, !llvm.ptr
      %319 = llvm.load %302 : !llvm.ptr -> i64
      %320 = arith.constant 1 : i32
      %322 = arith.extsi %320 : i32 to i64
      %321 = arith.addi %319, %322 : i64
      llvm.store %321, %302 : i64, !llvm.ptr
      cf.br ^bb54
    ^bb56:
    %323 = llvm.mlir.addressof @str_0 : !llvm.ptr
    %324 = llvm.load %296 : !llvm.ptr -> i64
    %325 = llvm.call @printf(%323, %324) vararg(!llvm.func<i32 (ptr, ...)>) : (!llvm.ptr, i64) -> i32
    func.call @free(%180) : (!llvm.ptr) -> ()
    func.call @free(%190) : (!llvm.ptr) -> ()
    %328 = arith.constant 0 : i32
    func.return %328 : i32
  }
}