Problem 201

Sum of uniquely-occurring 50-element subset sums of {1^2..100^2}.

View problem on Project Euler

Performance comparison

Flow source

# Project Euler 201
# Sum of uniquely-occurring 50-element subset sums of {1^2..100^2}.

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

function main() -> i32 {
    let N: i32 = 100
    let K: i32 = 50
    let mut max_sum: i64 = 0
    let mut i: i32 = N - K + 1
    while i <= N {
        max_sum = max_sum + (i as i64) * (i as i64)
        i = i + 1
    }
    let SZ: i64 = max_sum + 1
    # state[size][sum] capped at 2
    let state: ptr<i8> = calloc((K as i64 + 1) * SZ, 1)
    let low: ptr<i64> = calloc(K as i64 + 1, 8)
    let high: ptr<i64> = calloc(K as i64 + 1, 8)
    if state == null || low == null || high == null { return 1 }
    let MISSING: i64 = max_sum + 1
    i = 0
    while i <= K {
        low[i] = MISSING
        high[i] = 0 - 1
        i = i + 1
    }
    state[0] = 1
    low[0] = 0
    high[0] = 0

    let mut used: i32 = 0
    let mut sq_i: i32 = 1
    while sq_i <= N {
        let square: i64 = (sq_i as i64) * (sq_i as i64)
        let upper: i32 = used + 1
        let mut size: i32 = upper
        if size > K { size = K }
        while size > 0 {
            let lo: i64 = low[size - 1]
            if lo != MISSING {
                let hi: i64 = high[size - 1]
                let mut sub: i64 = hi
                while sub >= lo {
                    let ways: i32 = state[((size - 1) as i64) * SZ + sub] as i32
                    if ways != 0 {
                        let ns: i64 = sub + square
                        let mut nw: i32 = (state[(size as i64) * SZ + ns] as i32) + ways
                        if nw >= 2 { nw = 2 }
                        state[(size as i64) * SZ + ns] = nw as i8
                    }
                    sub = sub - 1
                }
                let nlo: i64 = lo + square
                let nhi: i64 = hi + square
                if nlo < low[size] { low[size] = nlo }
                if nhi > high[size] { high[size] = nhi }
            }
            size = size - 1
        }
        used = used + 1
        sq_i = sq_i + 1
    }

    let mut total: i64 = 0
    let mut s: i64 = low[K]
    while s <= high[K] {
        if state[(K as i64) * SZ + s] == 1 {
            total = total + s
        }
        s = s + 1
    }
    printf("%lld\n", total)
    free(state); free(low); free(high)
    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; }

int32_t main(void);



int32_t main(void) {
    int32_t N = 100;
    int32_t K = 50;
    int64_t max_sum = 0;
    int32_t i = ((N - K) + 1);
    while (i <= N) {
        max_sum = (max_sum + (((int64_t)(i)) * ((int64_t)(i))));
        i = (i + 1);
    }
    int64_t SZ = (max_sum + 1);
    int8_t* state = (int8_t*)(calloc(((((int64_t)(K)) + 1) * SZ), 1));
    int64_t* low = (int64_t*)(calloc((((int64_t)(K)) + 1), 8));
    int64_t* high = (int64_t*)(calloc((((int64_t)(K)) + 1), 8));
    if (((state == NULL || low == NULL) || high == NULL)) {
        return 1;
    }
    int64_t MISSING = (max_sum + 1);
    i = 0;
    while (i <= K) {
        low[i] = MISSING;
        high[i] = (0 - 1);
        i = (i + 1);
    }
    state[0] = 1;
    low[0] = 0;
    high[0] = 0;
    int32_t used = 0;
    int32_t sq_i = 1;
    while (sq_i <= N) {
        int64_t square = (((int64_t)(sq_i)) * ((int64_t)(sq_i)));
        int32_t upper = (used + 1);
        int32_t size = upper;
        if (size > K) {
            size = K;
        }
        while (size > 0) {
            int64_t lo = low[(size - 1)];
            if (lo != MISSING) {
                int64_t hi = high[(size - 1)];
                int64_t sub = hi;
                while (sub >= lo) {
                    int32_t ways = ((int32_t)(state[((((int64_t)((size - 1))) * SZ) + sub)]));
                    if (ways != 0) {
                        int64_t ns = (sub + square);
                        int32_t nw = (((int32_t)(state[((((int64_t)(size)) * SZ) + ns)])) + ways);
                        if (nw >= 2) {
                            nw = 2;
                        }
                        state[((((int64_t)(size)) * SZ) + ns)] = ((int8_t)(nw));
                    }
                    sub = (sub - 1);
                }
                int64_t nlo = (lo + square);
                int64_t nhi = (hi + square);
                if (nlo < low[size]) {
                    low[size] = nlo;
                }
                if (nhi > high[size]) {
                    high[size] = nhi;
                }
            }
            size = (size - 1);
        }
        used = (used + 1);
        sq_i = (sq_i + 1);
    }
    int64_t total = 0;
    int64_t s = low[K];
    while (s <= high[K]) {
        if (state[((((int64_t)(K)) * SZ) + s)] == 1) {
            total = (total + s);
        }
        s = (s + 1);
    }
    printf("%lld\n", total);
    free(state);
    free(low);
    free(high);
    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) -> ()
  func.func @main() -> i32 {
    %0 = arith.constant 100 : i32
    %1 = arith.constant 50 : i32
    %2 = arith.constant 0 : i32
    %3 = arith.extsi %2 : i32 to i64
    %4 = llvm.mlir.constant(1 : i64) : i64
    %5 = llvm.alloca %4 x i64 : (i64) -> !llvm.ptr
    llvm.store %3, %5 : i64, !llvm.ptr
    %6 = arith.subi %0, %1 : i32
    %7 = arith.constant 1 : i32
    %8 = arith.addi %6, %7 : i32
    %9 = llvm.mlir.constant(1 : i64) : i64
    %10 = llvm.alloca %9 x i32 : (i64) -> !llvm.ptr
    llvm.store %8, %10 : i32, !llvm.ptr
    cf.br ^bb0
    ^bb0:
    %11 = llvm.load %10 : !llvm.ptr -> i32
    %12 = arith.cmpi sle, %11, %0 : i32
    cf.cond_br %12, ^bb1, ^bb2
    ^bb1:
      %13 = llvm.load %5 : !llvm.ptr -> i64
      %14 = llvm.load %10 : !llvm.ptr -> i32
      %15 = arith.extsi %14 : i32 to i64
      %16 = llvm.load %10 : !llvm.ptr -> i32
      %17 = arith.extsi %16 : i32 to i64
      %18 = arith.muli %15, %17 : i64
      %19 = arith.addi %13, %18 : i64
      llvm.store %19, %5 : i64, !llvm.ptr
      %20 = llvm.load %10 : !llvm.ptr -> i32
      %21 = arith.constant 1 : i32
      %22 = arith.addi %20, %21 : i32
      llvm.store %22, %10 : i32, !llvm.ptr
      cf.br ^bb0
    ^bb2:
    %23 = llvm.load %5 : !llvm.ptr -> i64
    %24 = arith.constant 1 : i32
    %26 = arith.extsi %24 : i32 to i64
    %25 = arith.addi %23, %26 : i64
    %28 = arith.extsi %1 : i32 to i64
    %29 = arith.constant 1 : i32
    %31 = arith.extsi %29 : i32 to i64
    %30 = arith.addi %28, %31 : i64
    %32 = arith.muli %30, %25 : i64
    %33 = arith.constant 1 : i32
    %34 = arith.extsi %33 : i32 to i64
    %27 = func.call @calloc(%32, %34) : (i64, i64) -> !llvm.ptr
    %36 = arith.extsi %1 : i32 to i64
    %37 = arith.constant 1 : i32
    %39 = arith.extsi %37 : i32 to i64
    %38 = arith.addi %36, %39 : i64
    %40 = arith.constant 8 : i32
    %41 = arith.extsi %40 : i32 to i64
    %35 = func.call @calloc(%38, %41) : (i64, i64) -> !llvm.ptr
    %43 = arith.extsi %1 : i32 to i64
    %44 = arith.constant 1 : i32
    %46 = arith.extsi %44 : i32 to i64
    %45 = arith.addi %43, %46 : i64
    %47 = arith.constant 8 : i32
    %48 = arith.extsi %47 : i32 to i64
    %42 = func.call @calloc(%45, %48) : (i64, i64) -> !llvm.ptr
    %49 = llvm.mlir.zero : !llvm.ptr
    %50 = llvm.icmp "eq" %27, %49 : !llvm.ptr
    %51 = scf.if %50 -> (i1) {
      %52 = arith.constant true
      scf.yield %52 : i1
    } else {
      %53 = llvm.mlir.zero : !llvm.ptr
      %54 = llvm.icmp "eq" %35, %53 : !llvm.ptr
      scf.yield %54 : i1
    }
    %55 = scf.if %51 -> (i1) {
      %56 = arith.constant true
      scf.yield %56 : i1
    } else {
      %57 = llvm.mlir.zero : !llvm.ptr
      %58 = llvm.icmp "eq" %42, %57 : !llvm.ptr
      scf.yield %58 : i1
    }
    cf.cond_br %55, ^bb3, ^bb4
    ^bb3:
      %59 = arith.constant 1 : i32
      func.return %59 : i32
    ^bb4:
      cf.br ^bb5
    ^bb5:
    %60 = llvm.load %5 : !llvm.ptr -> i64
    %61 = arith.constant 1 : i32
    %63 = arith.extsi %61 : i32 to i64
    %62 = arith.addi %60, %63 : i64
    %64 = arith.constant 0 : i32
    llvm.store %64, %10 : i32, !llvm.ptr
    cf.br ^bb6
    ^bb6:
    %65 = llvm.load %10 : !llvm.ptr -> i32
    %66 = arith.cmpi sle, %65, %1 : i32
    cf.cond_br %66, ^bb7, ^bb8
    ^bb7:
      %67 = llvm.load %10 : !llvm.ptr -> i32
      %68 = arith.extsi %67 : i32 to i64
      %69 = llvm.getelementptr %35[%68] : (!llvm.ptr, i64) -> !llvm.ptr, i64
      llvm.store %62, %69 : i64, !llvm.ptr
      %70 = arith.constant 0 : i32
      %71 = arith.constant 1 : i32
      %72 = arith.subi %70, %71 : i32
      %73 = llvm.load %10 : !llvm.ptr -> i32
      %74 = arith.extsi %72 : i32 to i64
      %75 = arith.extsi %73 : i32 to i64
      %76 = llvm.getelementptr %42[%75] : (!llvm.ptr, i64) -> !llvm.ptr, i64
      llvm.store %74, %76 : i64, !llvm.ptr
      %77 = llvm.load %10 : !llvm.ptr -> i32
      %78 = arith.constant 1 : i32
      %79 = arith.addi %77, %78 : i32
      llvm.store %79, %10 : i32, !llvm.ptr
      cf.br ^bb6
    ^bb8:
    %80 = arith.constant 1 : i32
    %81 = arith.constant 0 : i32
    %82 = arith.trunci %80 : i32 to i8
    %83 = arith.extsi %81 : i32 to i64
    %84 = llvm.getelementptr %27[%83] : (!llvm.ptr, i64) -> !llvm.ptr, i8
    llvm.store %82, %84 : i8, !llvm.ptr
    %85 = arith.constant 0 : i32
    %86 = arith.constant 0 : i32
    %87 = arith.extsi %85 : i32 to i64
    %88 = arith.extsi %86 : i32 to i64
    %89 = llvm.getelementptr %35[%88] : (!llvm.ptr, i64) -> !llvm.ptr, i64
    llvm.store %87, %89 : i64, !llvm.ptr
    %90 = arith.constant 0 : i32
    %91 = arith.constant 0 : i32
    %92 = arith.extsi %90 : i32 to i64
    %93 = arith.extsi %91 : i32 to i64
    %94 = llvm.getelementptr %42[%93] : (!llvm.ptr, i64) -> !llvm.ptr, i64
    llvm.store %92, %94 : i64, !llvm.ptr
    %95 = arith.constant 0 : i32
    %96 = llvm.mlir.constant(1 : i64) : i64
    %97 = llvm.alloca %96 x i32 : (i64) -> !llvm.ptr
    llvm.store %95, %97 : i32, !llvm.ptr
    %98 = arith.constant 1 : i32
    %99 = llvm.mlir.constant(1 : i64) : i64
    %100 = llvm.alloca %99 x i32 : (i64) -> !llvm.ptr
    llvm.store %98, %100 : i32, !llvm.ptr
    cf.br ^bb9
    ^bb9:
    %101 = llvm.load %100 : !llvm.ptr -> i32
    %102 = arith.cmpi sle, %101, %0 : i32
    cf.cond_br %102, ^bb10, ^bb11
    ^bb10:
      %103 = llvm.load %100 : !llvm.ptr -> i32
      %104 = arith.extsi %103 : i32 to i64
      %105 = llvm.load %100 : !llvm.ptr -> i32
      %106 = arith.extsi %105 : i32 to i64
      %107 = arith.muli %104, %106 : i64
      %108 = llvm.load %97 : !llvm.ptr -> i32
      %109 = arith.constant 1 : i32
      %110 = arith.addi %108, %109 : i32
      %111 = llvm.mlir.constant(1 : i64) : i64
      %112 = llvm.alloca %111 x i32 : (i64) -> !llvm.ptr
      llvm.store %110, %112 : i32, !llvm.ptr
      %113 = llvm.load %112 : !llvm.ptr -> i32
      %114 = arith.cmpi sgt, %113, %1 : i32
      cf.cond_br %114, ^bb12, ^bb13
      ^bb12:
        llvm.store %1, %112 : i32, !llvm.ptr
        cf.br ^bb14
      ^bb13:
        cf.br ^bb14
      ^bb14:
      cf.br ^bb15
      ^bb15:
      %115 = llvm.load %112 : !llvm.ptr -> i32
      %116 = arith.constant 0 : i32
      %117 = arith.cmpi sgt, %115, %116 : i32
      cf.cond_br %117, ^bb16, ^bb17
      ^bb16:
        %119 = llvm.load %112 : !llvm.ptr -> i32
        %120 = arith.constant 1 : i32
        %121 = arith.subi %119, %120 : i32
        %122 = arith.extsi %121 : i32 to i64
        %123 = llvm.getelementptr %35[%122] : (!llvm.ptr, i64) -> !llvm.ptr, i64
        %118 = llvm.load %123 : !llvm.ptr -> i64
        %124 = arith.cmpi ne, %118, %62 : i64
        cf.cond_br %124, ^bb18, ^bb19
        ^bb18:
          %126 = llvm.load %112 : !llvm.ptr -> i32
          %127 = arith.constant 1 : i32
          %128 = arith.subi %126, %127 : i32
          %129 = arith.extsi %128 : i32 to i64
          %130 = llvm.getelementptr %42[%129] : (!llvm.ptr, i64) -> !llvm.ptr, i64
          %125 = llvm.load %130 : !llvm.ptr -> i64
          %131 = llvm.mlir.constant(1 : i64) : i64
          %132 = llvm.alloca %131 x i64 : (i64) -> !llvm.ptr
          llvm.store %125, %132 : i64, !llvm.ptr
          cf.br ^bb21
          ^bb21:
          %133 = llvm.load %132 : !llvm.ptr -> i64
          %134 = arith.cmpi sge, %133, %118 : i64
          cf.cond_br %134, ^bb22, ^bb23
          ^bb22:
            %136 = llvm.load %112 : !llvm.ptr -> i32
            %137 = arith.constant 1 : i32
            %138 = arith.subi %136, %137 : i32
            %139 = arith.extsi %138 : i32 to i64
            %140 = arith.muli %139, %25 : i64
            %141 = llvm.load %132 : !llvm.ptr -> i64
            %142 = arith.addi %140, %141 : i64
            %143 = llvm.getelementptr %27[%142] : (!llvm.ptr, i64) -> !llvm.ptr, i8
            %135 = llvm.load %143 : !llvm.ptr -> i8
            %144 = arith.extsi %135 : i8 to i32
            %145 = arith.constant 0 : i32
            %146 = arith.cmpi ne, %144, %145 : i32
            cf.cond_br %146, ^bb24, ^bb25
            ^bb24:
              %147 = llvm.load %132 : !llvm.ptr -> i64
              %148 = arith.addi %147, %107 : i64
              %150 = llvm.load %112 : !llvm.ptr -> i32
              %151 = arith.extsi %150 : i32 to i64
              %152 = arith.muli %151, %25 : i64
              %153 = arith.addi %152, %148 : i64
              %154 = llvm.getelementptr %27[%153] : (!llvm.ptr, i64) -> !llvm.ptr, i8
              %149 = llvm.load %154 : !llvm.ptr -> i8
              %155 = arith.extsi %149 : i8 to i32
              %156 = arith.addi %155, %144 : i32
              %157 = llvm.mlir.constant(1 : i64) : i64
              %158 = llvm.alloca %157 x i32 : (i64) -> !llvm.ptr
              llvm.store %156, %158 : i32, !llvm.ptr
              %159 = llvm.load %158 : !llvm.ptr -> i32
              %160 = arith.constant 2 : i32
              %161 = arith.cmpi sge, %159, %160 : i32
              cf.cond_br %161, ^bb27, ^bb28
              ^bb27:
                %162 = arith.constant 2 : i32
                llvm.store %162, %158 : i32, !llvm.ptr
                cf.br ^bb29
              ^bb28:
                cf.br ^bb29
              ^bb29:
              %163 = llvm.load %158 : !llvm.ptr -> i32
              %164 = arith.trunci %163 : i32 to i8
              %165 = llvm.load %112 : !llvm.ptr -> i32
              %166 = arith.extsi %165 : i32 to i64
              %167 = arith.muli %166, %25 : i64
              %168 = arith.addi %167, %148 : i64
              %169 = llvm.getelementptr %27[%168] : (!llvm.ptr, i64) -> !llvm.ptr, i8
              llvm.store %164, %169 : i8, !llvm.ptr
              cf.br ^bb26
            ^bb25:
              cf.br ^bb26
            ^bb26:
            %170 = llvm.load %132 : !llvm.ptr -> i64
            %171 = arith.constant 1 : i32
            %173 = arith.extsi %171 : i32 to i64
            %172 = arith.subi %170, %173 : i64
            llvm.store %172, %132 : i64, !llvm.ptr
            cf.br ^bb21
          ^bb23:
          %174 = arith.addi %118, %107 : i64
          %175 = arith.addi %125, %107 : i64
          %177 = llvm.load %112 : !llvm.ptr -> i32
          %178 = arith.extsi %177 : i32 to i64
          %179 = llvm.getelementptr %35[%178] : (!llvm.ptr, i64) -> !llvm.ptr, i64
          %176 = llvm.load %179 : !llvm.ptr -> i64
          %180 = arith.cmpi slt, %174, %176 : i64
          cf.cond_br %180, ^bb30, ^bb31
          ^bb30:
            %181 = llvm.load %112 : !llvm.ptr -> i32
            %182 = arith.extsi %181 : i32 to i64
            %183 = llvm.getelementptr %35[%182] : (!llvm.ptr, i64) -> !llvm.ptr, i64
            llvm.store %174, %183 : i64, !llvm.ptr
            cf.br ^bb32
          ^bb31:
            cf.br ^bb32
          ^bb32:
          %185 = llvm.load %112 : !llvm.ptr -> i32
          %186 = arith.extsi %185 : i32 to i64
          %187 = llvm.getelementptr %42[%186] : (!llvm.ptr, i64) -> !llvm.ptr, i64
          %184 = llvm.load %187 : !llvm.ptr -> i64
          %188 = arith.cmpi sgt, %175, %184 : i64
          cf.cond_br %188, ^bb33, ^bb34
          ^bb33:
            %189 = llvm.load %112 : !llvm.ptr -> i32
            %190 = arith.extsi %189 : i32 to i64
            %191 = llvm.getelementptr %42[%190] : (!llvm.ptr, i64) -> !llvm.ptr, i64
            llvm.store %175, %191 : i64, !llvm.ptr
            cf.br ^bb35
          ^bb34:
            cf.br ^bb35
          ^bb35:
          cf.br ^bb20
        ^bb19:
          cf.br ^bb20
        ^bb20:
        %192 = llvm.load %112 : !llvm.ptr -> i32
        %193 = arith.constant 1 : i32
        %194 = arith.subi %192, %193 : i32
        llvm.store %194, %112 : i32, !llvm.ptr
        cf.br ^bb15
      ^bb17:
      %195 = llvm.load %97 : !llvm.ptr -> i32
      %196 = arith.constant 1 : i32
      %197 = arith.addi %195, %196 : i32
      llvm.store %197, %97 : i32, !llvm.ptr
      %198 = llvm.load %100 : !llvm.ptr -> i32
      %199 = arith.constant 1 : i32
      %200 = arith.addi %198, %199 : i32
      llvm.store %200, %100 : i32, !llvm.ptr
      cf.br ^bb9
    ^bb11:
    %201 = arith.constant 0 : i32
    %202 = arith.extsi %201 : i32 to i64
    %203 = llvm.mlir.constant(1 : i64) : i64
    %204 = llvm.alloca %203 x i64 : (i64) -> !llvm.ptr
    llvm.store %202, %204 : i64, !llvm.ptr
    %206 = arith.extsi %1 : i32 to i64
    %207 = llvm.getelementptr %35[%206] : (!llvm.ptr, i64) -> !llvm.ptr, i64
    %205 = llvm.load %207 : !llvm.ptr -> i64
    %208 = llvm.mlir.constant(1 : i64) : i64
    %209 = llvm.alloca %208 x i64 : (i64) -> !llvm.ptr
    llvm.store %205, %209 : i64, !llvm.ptr
    cf.br ^bb36
    ^bb36:
    %210 = llvm.load %209 : !llvm.ptr -> i64
    %212 = arith.extsi %1 : i32 to i64
    %213 = llvm.getelementptr %42[%212] : (!llvm.ptr, i64) -> !llvm.ptr, i64
    %211 = llvm.load %213 : !llvm.ptr -> i64
    %214 = arith.cmpi sle, %210, %211 : i64
    cf.cond_br %214, ^bb37, ^bb38
    ^bb37:
      %216 = arith.extsi %1 : i32 to i64
      %217 = arith.muli %216, %25 : i64
      %218 = llvm.load %209 : !llvm.ptr -> i64
      %219 = arith.addi %217, %218 : i64
      %220 = llvm.getelementptr %27[%219] : (!llvm.ptr, i64) -> !llvm.ptr, i8
      %215 = llvm.load %220 : !llvm.ptr -> i8
      %221 = arith.constant 1 : i32
      %223 = arith.extsi %215 : i8 to i32
      %222 = arith.cmpi eq, %223, %221 : i32
      cf.cond_br %222, ^bb39, ^bb40
      ^bb39:
        %224 = llvm.load %204 : !llvm.ptr -> i64
        %225 = llvm.load %209 : !llvm.ptr -> i64
        %226 = arith.addi %224, %225 : i64
        llvm.store %226, %204 : i64, !llvm.ptr
        cf.br ^bb41
      ^bb40:
        cf.br ^bb41
      ^bb41:
      %227 = llvm.load %209 : !llvm.ptr -> i64
      %228 = arith.constant 1 : i32
      %230 = arith.extsi %228 : i32 to i64
      %229 = arith.addi %227, %230 : i64
      llvm.store %229, %209 : i64, !llvm.ptr
      cf.br ^bb36
    ^bb38:
    %231 = llvm.mlir.addressof @str_0 : !llvm.ptr
    %232 = llvm.load %204 : !llvm.ptr -> i64
    %233 = llvm.call @printf(%231, %232) vararg(!llvm.func<i32 (ptr, ...)>) : (!llvm.ptr, i64) -> i32
    func.call @free(%27) : (!llvm.ptr) -> ()
    func.call @free(%35) : (!llvm.ptr) -> ()
    func.call @free(%42) : (!llvm.ptr) -> ()
    %237 = arith.constant 0 : i32
    func.return %237 : i32
  }
}