Problem 315

Sam vs Max digital-root clocks for primes in [10^7, 2*10^7].

View problem on Project Euler

Performance comparison

Flow source

# Project Euler 315
# Sam vs Max digital-root clocks for primes in [10^7, 2*10^7].

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

function popcnt(x0: i64) -> i32 {
    let mut x: i64 = x0
    let mut r: i32 = 0
    while x > 0 {
        x = x & (x - 1)
        r = r + 1
    }
    return r
}

function digit_sum(x0: i64) -> i64 {
    let mut x: i64 = x0
    let mut r: i64 = 0
    while x > 0 {
        r = r + x % 10
        x = x / 10
    }
    return r
}

function get_segments(x0: i64, seg: ptr<i64>) -> i64 {
    let mut x: i64 = x0
    let mut result: i64 = 0
    let mut shift: i64 = 0
    while x > 0 {
        result = result | (seg[x % 10] << shift)
        x = x / 10
        shift = shift + 8
    }
    return result
}

function sam(x0: i64, seg: ptr<i64>, cache: ptr<i32>) -> i32 {
    let mut x: i64 = x0
    if x < 100 && cache[x] > 0 { return cache[x] }
    let segments: i64 = get_segments(x, seg)
    let mut result: i32 = 2 * popcnt(segments)
    if x > 9 {
        result = result + sam(digit_sum(x), seg, cache)
    }
    if x < 100 { cache[x] = result }
    return result
}

function max_clock(x0: i64, prev: i64, seg: ptr<i64>) -> i32 {
    let segments: i64 = get_segments(x0, seg)
    let mut result: i32 = popcnt(segments ^ prev)
    if x0 > 9 {
        result = result + max_clock(digit_sum(x0), segments, seg)
    } else {
        result = result + popcnt(segments)
    }
    return result
}

function main() -> i32 {
    let lo: i64 = 10000000
    let hi: i64 = 20000000
    let half: i64 = (hi >> 1) + 1
    let sieve: ptr<i8> = calloc(half, 1)
    let seg: ptr<i64> = calloc(10, 8)
    let cache: ptr<i32> = calloc(100, 4)
    if sieve == null || seg == null || cache == null { return 1 }
    # segment bitmasks
    seg[0] = 119; seg[1] = 36; seg[2] = 93; seg[3] = 109; seg[4] = 46
    seg[5] = 107; seg[6] = 123; seg[7] = 39; seg[8] = 127; seg[9] = 111

    for i in 0..half { sieve[i] = 1 }
    sieve[0] = 0
    let mut i: i64 = 1
    while 2 * i * i < half {
        if sieve[i] == 1 {
            let mut current: i64 = 3 * i + 1
            let step: i64 = 2 * i + 1
            while current < half {
                sieve[current] = 0
                current = current + step
            }
        }
        i = i + 1
    }

    let mut sum_sam: i64 = 0
    let mut sum_max: i64 = 0
    let mut p: i64 = lo
    if p % 2 == 0 { p = p + 1 }
    while p <= hi {
        if sieve[p >> 1] == 1 {
            sum_sam = sum_sam + (sam(p, seg, cache) as i64)
            sum_max = sum_max + (max_clock(p, 0, seg) as i64)
        }
        p = p + 2
    }
    printf("%lld\n", sum_sam - sum_max)
    free(sieve); free(seg); free(cache)
    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 popcnt_i64(int64_t x0);
int64_t digit_sum_i64(int64_t x0);
int64_t get_segments_i64_ptr_i64(int64_t x0, int64_t* seg);
int32_t sam_i64_ptr_i64_ptr_i32(int64_t x0, int64_t* seg, int32_t* cache);
int32_t max_clock_i64_i64_ptr_i64(int64_t x0, int64_t prev, int64_t* seg);
int32_t main(void);



int32_t popcnt_i64(int64_t x0) {
    int64_t x = x0;
    int32_t r = 0;
    while (x > 0) {
        x = (x & (x - 1));
        r = (r + 1);
    }
    return r;
}

int64_t digit_sum_i64(int64_t x0) {
    int64_t x = x0;
    int64_t r = 0;
    while (x > 0) {
        r = (r + FLOW_CHECKED_MOD((x), (10)));
        x = FLOW_CHECKED_DIV((x), (10));
    }
    return r;
}

int64_t get_segments_i64_ptr_i64(int64_t x0, int64_t* seg) {
    int64_t x = x0;
    int64_t result = 0;
    int64_t shift = 0;
    while (x > 0) {
        result = (result | FLOW_CHECKED_SHL((seg[FLOW_CHECKED_MOD((x), (10))]), (shift)));
        x = FLOW_CHECKED_DIV((x), (10));
        shift = (shift + 8);
    }
    return result;
}

int32_t sam_i64_ptr_i64_ptr_i32(int64_t x0, int64_t* seg, int32_t* cache) {
    int64_t x = x0;
    if ((x < 100 && cache[x] > 0)) {
        return cache[x];
    }
    int64_t segments = get_segments_i64_ptr_i64(x, seg);
    int32_t result = (2 * popcnt_i64(segments));
    if (x > 9) {
        result = (result + sam_i64_ptr_i64_ptr_i32(digit_sum_i64(x), seg, cache));
    }
    if (x < 100) {
        cache[x] = result;
    }
    return result;
}

int32_t max_clock_i64_i64_ptr_i64(int64_t x0, int64_t prev, int64_t* seg) {
    int64_t segments = get_segments_i64_ptr_i64(x0, seg);
    int32_t result = popcnt_i64((segments ^ prev));
    if (x0 > 9) {
        result = (result + max_clock_i64_i64_ptr_i64(digit_sum_i64(x0), segments, seg));
    } else {
        result = (result + popcnt_i64(segments));
    }
    return result;
}

int32_t main(void) {
    int64_t lo = 10000000;
    int64_t hi = 20000000;
    int64_t half = (FLOW_CHECKED_SHR((hi), (1)) + 1);
    int8_t* sieve = (int8_t*)(calloc(half, 1));
    int64_t* seg = (int64_t*)(calloc(10, 8));
    int32_t* cache = (int32_t*)(calloc(100, 4));
    if (((sieve == NULL || seg == NULL) || cache == NULL)) {
        return 1;
    }
    seg[0] = 119;
    seg[1] = 36;
    seg[2] = 93;
    seg[3] = 109;
    seg[4] = 46;
    seg[5] = 107;
    seg[6] = 123;
    seg[7] = 39;
    seg[8] = 127;
    seg[9] = 111;
    int32_t __flow_step_1 = 1;
    for (int32_t i = 0; (0 <= half) ? i < half : i > half; i += (0 <= half) ? 1 : -1) {
        sieve[i] = 1;
    }
    sieve[0] = 0;
    int64_t i = 1;
    while (((2 * i) * i) < half) {
        if (sieve[i] == 1) {
            int64_t current = ((3 * i) + 1);
            int64_t step = ((2 * i) + 1);
            while (current < half) {
                sieve[current] = 0;
                current = (current + step);
            }
        }
        i = (i + 1);
    }
    int64_t sum_sam = 0;
    int64_t sum_max = 0;
    int64_t p = lo;
    if (FLOW_CHECKED_MOD((p), (2)) == 0) {
        p = (p + 1);
    }
    while (p <= hi) {
        if (sieve[FLOW_CHECKED_SHR((p), (1))] == 1) {
            sum_sam = (sum_sam + ((int64_t)(sam_i64_ptr_i64_ptr_i32(p, seg, cache))));
            sum_max = (sum_max + ((int64_t)(max_clock_i64_i64_ptr_i64(p, 0, seg))));
        }
        p = (p + 2);
    }
    printf("%lld\n", (sum_sam - sum_max));
    free(sieve);
    free(seg);
    free(cache);
    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 @popcnt(%arg0: i64) -> i32 {
    %0 = llvm.mlir.constant(1 : i64) : i64
    %1 = llvm.alloca %0 x i64 : (i64) -> !llvm.ptr
    llvm.store %arg0, %1 : i64, !llvm.ptr
    %2 = arith.constant 0 : i32
    %3 = llvm.mlir.constant(1 : i64) : i64
    %4 = llvm.alloca %3 x i32 : (i64) -> !llvm.ptr
    llvm.store %2, %4 : i32, !llvm.ptr
    cf.br ^bb0
    ^bb0:
    %5 = llvm.load %1 : !llvm.ptr -> i64
    %6 = arith.constant 0 : i32
    %8 = arith.extsi %6 : i32 to i64
    %7 = arith.cmpi sgt, %5, %8 : i64
    cf.cond_br %7, ^bb1, ^bb2
    ^bb1:
      %9 = llvm.load %1 : !llvm.ptr -> i64
      %10 = llvm.load %1 : !llvm.ptr -> i64
      %11 = arith.constant 1 : i32
      %13 = arith.extsi %11 : i32 to i64
      %12 = arith.subi %10, %13 : i64
      %14 = arith.andi %9, %12 : i64
      llvm.store %14, %1 : i64, !llvm.ptr
      %15 = llvm.load %4 : !llvm.ptr -> i32
      %16 = arith.constant 1 : i32
      %17 = arith.addi %15, %16 : i32
      llvm.store %17, %4 : i32, !llvm.ptr
      cf.br ^bb0
    ^bb2:
    %18 = llvm.load %4 : !llvm.ptr -> i32
    func.return %18 : i32
  }
  func.func @digit_sum(%arg0: i64) -> i64 {
    %19 = llvm.mlir.constant(1 : i64) : i64
    %20 = llvm.alloca %19 x i64 : (i64) -> !llvm.ptr
    llvm.store %arg0, %20 : i64, !llvm.ptr
    %21 = arith.constant 0 : i32
    %22 = arith.extsi %21 : i32 to i64
    %23 = llvm.mlir.constant(1 : i64) : i64
    %24 = llvm.alloca %23 x i64 : (i64) -> !llvm.ptr
    llvm.store %22, %24 : i64, !llvm.ptr
    cf.br ^bb3
    ^bb3:
    %25 = llvm.load %20 : !llvm.ptr -> i64
    %26 = arith.constant 0 : i32
    %28 = arith.extsi %26 : i32 to i64
    %27 = arith.cmpi sgt, %25, %28 : i64
    cf.cond_br %27, ^bb4, ^bb5
    ^bb4:
      %29 = llvm.load %24 : !llvm.ptr -> i64
      %30 = llvm.load %20 : !llvm.ptr -> i64
      %31 = arith.constant 10 : i32
      %33 = arith.extsi %31 : i32 to i64
      %32 = arith.remsi %30, %33 : i64
      %34 = arith.addi %29, %32 : i64
      llvm.store %34, %24 : i64, !llvm.ptr
      %35 = llvm.load %20 : !llvm.ptr -> i64
      %36 = arith.constant 10 : i32
      %38 = arith.extsi %36 : i32 to i64
      %37 = arith.divsi %35, %38 : i64
      llvm.store %37, %20 : i64, !llvm.ptr
      cf.br ^bb3
    ^bb5:
    %39 = llvm.load %24 : !llvm.ptr -> i64
    func.return %39 : i64
  }
  func.func @get_segments(%arg0: i64, %arg1: !llvm.ptr) -> i64 {
    %40 = llvm.mlir.constant(1 : i64) : i64
    %41 = llvm.alloca %40 x i64 : (i64) -> !llvm.ptr
    llvm.store %arg0, %41 : i64, !llvm.ptr
    %42 = arith.constant 0 : i32
    %43 = arith.extsi %42 : i32 to i64
    %44 = llvm.mlir.constant(1 : i64) : i64
    %45 = llvm.alloca %44 x i64 : (i64) -> !llvm.ptr
    llvm.store %43, %45 : i64, !llvm.ptr
    %46 = arith.constant 0 : i32
    %47 = arith.extsi %46 : i32 to i64
    %48 = llvm.mlir.constant(1 : i64) : i64
    %49 = llvm.alloca %48 x i64 : (i64) -> !llvm.ptr
    llvm.store %47, %49 : i64, !llvm.ptr
    cf.br ^bb6
    ^bb6:
    %50 = llvm.load %41 : !llvm.ptr -> i64
    %51 = arith.constant 0 : i32
    %53 = arith.extsi %51 : i32 to i64
    %52 = arith.cmpi sgt, %50, %53 : i64
    cf.cond_br %52, ^bb7, ^bb8
    ^bb7:
      %54 = llvm.load %45 : !llvm.ptr -> i64
      %56 = llvm.load %41 : !llvm.ptr -> i64
      %57 = arith.constant 10 : i32
      %59 = arith.extsi %57 : i32 to i64
      %58 = arith.remsi %56, %59 : i64
      %60 = llvm.getelementptr %arg1[%58] : (!llvm.ptr, i64) -> !llvm.ptr, i64
      %55 = llvm.load %60 : !llvm.ptr -> i64
      %61 = llvm.load %49 : !llvm.ptr -> i64
      %62 = arith.shli %55, %61 : i64
      %63 = arith.ori %54, %62 : i64
      llvm.store %63, %45 : i64, !llvm.ptr
      %64 = llvm.load %41 : !llvm.ptr -> i64
      %65 = arith.constant 10 : i32
      %67 = arith.extsi %65 : i32 to i64
      %66 = arith.divsi %64, %67 : i64
      llvm.store %66, %41 : i64, !llvm.ptr
      %68 = llvm.load %49 : !llvm.ptr -> i64
      %69 = arith.constant 8 : i32
      %71 = arith.extsi %69 : i32 to i64
      %70 = arith.addi %68, %71 : i64
      llvm.store %70, %49 : i64, !llvm.ptr
      cf.br ^bb6
    ^bb8:
    %72 = llvm.load %45 : !llvm.ptr -> i64
    func.return %72 : i64
  }
  func.func @sam(%arg0: i64, %arg1: !llvm.ptr, %arg2: !llvm.ptr) -> i32 {
    %73 = llvm.mlir.constant(1 : i64) : i64
    %74 = llvm.alloca %73 x i64 : (i64) -> !llvm.ptr
    llvm.store %arg0, %74 : i64, !llvm.ptr
    %75 = llvm.load %74 : !llvm.ptr -> i64
    %76 = arith.constant 100 : i32
    %78 = arith.extsi %76 : i32 to i64
    %77 = arith.cmpi slt, %75, %78 : i64
    %79 = scf.if %77 -> (i1) {
      %81 = llvm.load %74 : !llvm.ptr -> i64
      %82 = llvm.getelementptr %arg2[%81] : (!llvm.ptr, i64) -> !llvm.ptr, i32
      %80 = llvm.load %82 : !llvm.ptr -> i32
      %83 = arith.constant 0 : i32
      %84 = arith.cmpi sgt, %80, %83 : i32
      scf.yield %84 : i1
    } else {
      %85 = arith.constant false
      scf.yield %85 : i1
    }
    cf.cond_br %79, ^bb9, ^bb10
    ^bb9:
      %87 = llvm.load %74 : !llvm.ptr -> i64
      %88 = llvm.getelementptr %arg2[%87] : (!llvm.ptr, i64) -> !llvm.ptr, i32
      %86 = llvm.load %88 : !llvm.ptr -> i32
      func.return %86 : i32
    ^bb10:
      cf.br ^bb11
    ^bb11:
    %90 = llvm.load %74 : !llvm.ptr -> i64
    %89 = func.call @get_segments(%90, %arg1) : (i64, !llvm.ptr) -> i64
    %91 = arith.constant 2 : i32
    %92 = func.call @popcnt(%89) : (i64) -> i32
    %93 = arith.muli %91, %92 : i32
    %94 = llvm.mlir.constant(1 : i64) : i64
    %95 = llvm.alloca %94 x i32 : (i64) -> !llvm.ptr
    llvm.store %93, %95 : i32, !llvm.ptr
    %96 = llvm.load %74 : !llvm.ptr -> i64
    %97 = arith.constant 9 : i32
    %99 = arith.extsi %97 : i32 to i64
    %98 = arith.cmpi sgt, %96, %99 : i64
    cf.cond_br %98, ^bb12, ^bb13
    ^bb12:
      %100 = llvm.load %95 : !llvm.ptr -> i32
      %103 = llvm.load %74 : !llvm.ptr -> i64
      %102 = func.call @digit_sum(%103) : (i64) -> i64
      %101 = func.call @sam(%102, %arg1, %arg2) : (i64, !llvm.ptr, !llvm.ptr) -> i32
      %104 = arith.addi %100, %101 : i32
      llvm.store %104, %95 : i32, !llvm.ptr
      cf.br ^bb14
    ^bb13:
      cf.br ^bb14
    ^bb14:
    %105 = llvm.load %74 : !llvm.ptr -> i64
    %106 = arith.constant 100 : i32
    %108 = arith.extsi %106 : i32 to i64
    %107 = arith.cmpi slt, %105, %108 : i64
    cf.cond_br %107, ^bb15, ^bb16
    ^bb15:
      %109 = llvm.load %95 : !llvm.ptr -> i32
      %110 = llvm.load %74 : !llvm.ptr -> i64
      %111 = llvm.getelementptr %arg2[%110] : (!llvm.ptr, i64) -> !llvm.ptr, i32
      llvm.store %109, %111 : i32, !llvm.ptr
      cf.br ^bb17
    ^bb16:
      cf.br ^bb17
    ^bb17:
    %112 = llvm.load %95 : !llvm.ptr -> i32
    func.return %112 : i32
  }
  func.func @max_clock(%arg0: i64, %arg1: i64, %arg2: !llvm.ptr) -> i32 {
    %113 = func.call @get_segments(%arg0, %arg2) : (i64, !llvm.ptr) -> i64
    %115 = arith.xori %113, %arg1 : i64
    %114 = func.call @popcnt(%115) : (i64) -> i32
    %116 = llvm.mlir.constant(1 : i64) : i64
    %117 = llvm.alloca %116 x i32 : (i64) -> !llvm.ptr
    llvm.store %114, %117 : i32, !llvm.ptr
    %118 = arith.constant 9 : i32
    %120 = arith.extsi %118 : i32 to i64
    %119 = arith.cmpi sgt, %arg0, %120 : i64
    cf.cond_br %119, ^bb18, ^bb19
    ^bb18:
      %121 = llvm.load %117 : !llvm.ptr -> i32
      %123 = func.call @digit_sum(%arg0) : (i64) -> i64
      %122 = func.call @max_clock(%123, %113, %arg2) : (i64, i64, !llvm.ptr) -> i32
      %124 = arith.addi %121, %122 : i32
      llvm.store %124, %117 : i32, !llvm.ptr
      cf.br ^bb20
    ^bb19:
      %125 = llvm.load %117 : !llvm.ptr -> i32
      %126 = func.call @popcnt(%113) : (i64) -> i32
      %127 = arith.addi %125, %126 : i32
      llvm.store %127, %117 : i32, !llvm.ptr
      cf.br ^bb20
    ^bb20:
    %128 = llvm.load %117 : !llvm.ptr -> i32
    func.return %128 : i32
  }
  func.func @main() -> i32 {
    %129 = arith.constant 10000000 : i32
    %130 = arith.extsi %129 : i32 to i64
    %131 = arith.constant 20000000 : i32
    %132 = arith.extsi %131 : i32 to i64
    %133 = arith.constant 1 : i32
    %135 = arith.extsi %133 : i32 to i64
    %134 = arith.shrsi %132, %135 : i64
    %136 = arith.constant 1 : i32
    %138 = arith.extsi %136 : i32 to i64
    %137 = arith.addi %134, %138 : i64
    %140 = arith.constant 1 : i32
    %141 = arith.extsi %140 : i32 to i64
    %139 = func.call @calloc(%137, %141) : (i64, i64) -> !llvm.ptr
    %143 = arith.constant 10 : i32
    %144 = arith.constant 8 : i32
    %145 = arith.extsi %143 : i32 to i64
    %146 = arith.extsi %144 : i32 to i64
    %142 = func.call @calloc(%145, %146) : (i64, i64) -> !llvm.ptr
    %148 = arith.constant 100 : i32
    %149 = arith.constant 4 : i32
    %150 = arith.extsi %148 : i32 to i64
    %151 = arith.extsi %149 : i32 to i64
    %147 = func.call @calloc(%150, %151) : (i64, i64) -> !llvm.ptr
    %152 = llvm.mlir.zero : !llvm.ptr
    %153 = llvm.icmp "eq" %139, %152 : !llvm.ptr
    %154 = scf.if %153 -> (i1) {
      %155 = arith.constant true
      scf.yield %155 : i1
    } else {
      %156 = llvm.mlir.zero : !llvm.ptr
      %157 = llvm.icmp "eq" %142, %156 : !llvm.ptr
      scf.yield %157 : i1
    }
    %158 = scf.if %154 -> (i1) {
      %159 = arith.constant true
      scf.yield %159 : i1
    } else {
      %160 = llvm.mlir.zero : !llvm.ptr
      %161 = llvm.icmp "eq" %147, %160 : !llvm.ptr
      scf.yield %161 : i1
    }
    cf.cond_br %158, ^bb21, ^bb22
    ^bb21:
      %162 = arith.constant 1 : i32
      func.return %162 : i32
    ^bb22:
      cf.br ^bb23
    ^bb23:
    %163 = arith.constant 119 : i32
    %164 = arith.constant 0 : i32
    %165 = arith.extsi %163 : i32 to i64
    %166 = arith.extsi %164 : i32 to i64
    %167 = llvm.getelementptr %142[%166] : (!llvm.ptr, i64) -> !llvm.ptr, i64
    llvm.store %165, %167 : i64, !llvm.ptr
    %168 = arith.constant 36 : i32
    %169 = arith.constant 1 : i32
    %170 = arith.extsi %168 : i32 to i64
    %171 = arith.extsi %169 : i32 to i64
    %172 = llvm.getelementptr %142[%171] : (!llvm.ptr, i64) -> !llvm.ptr, i64
    llvm.store %170, %172 : i64, !llvm.ptr
    %173 = arith.constant 93 : i32
    %174 = arith.constant 2 : i32
    %175 = arith.extsi %173 : i32 to i64
    %176 = arith.extsi %174 : i32 to i64
    %177 = llvm.getelementptr %142[%176] : (!llvm.ptr, i64) -> !llvm.ptr, i64
    llvm.store %175, %177 : i64, !llvm.ptr
    %178 = arith.constant 109 : i32
    %179 = arith.constant 3 : i32
    %180 = arith.extsi %178 : i32 to i64
    %181 = arith.extsi %179 : i32 to i64
    %182 = llvm.getelementptr %142[%181] : (!llvm.ptr, i64) -> !llvm.ptr, i64
    llvm.store %180, %182 : i64, !llvm.ptr
    %183 = arith.constant 46 : i32
    %184 = arith.constant 4 : i32
    %185 = arith.extsi %183 : i32 to i64
    %186 = arith.extsi %184 : i32 to i64
    %187 = llvm.getelementptr %142[%186] : (!llvm.ptr, i64) -> !llvm.ptr, i64
    llvm.store %185, %187 : i64, !llvm.ptr
    %188 = arith.constant 107 : i32
    %189 = arith.constant 5 : i32
    %190 = arith.extsi %188 : i32 to i64
    %191 = arith.extsi %189 : i32 to i64
    %192 = llvm.getelementptr %142[%191] : (!llvm.ptr, i64) -> !llvm.ptr, i64
    llvm.store %190, %192 : i64, !llvm.ptr
    %193 = arith.constant 123 : i32
    %194 = arith.constant 6 : i32
    %195 = arith.extsi %193 : i32 to i64
    %196 = arith.extsi %194 : i32 to i64
    %197 = llvm.getelementptr %142[%196] : (!llvm.ptr, i64) -> !llvm.ptr, i64
    llvm.store %195, %197 : i64, !llvm.ptr
    %198 = arith.constant 39 : i32
    %199 = arith.constant 7 : i32
    %200 = arith.extsi %198 : i32 to i64
    %201 = arith.extsi %199 : i32 to i64
    %202 = llvm.getelementptr %142[%201] : (!llvm.ptr, i64) -> !llvm.ptr, i64
    llvm.store %200, %202 : i64, !llvm.ptr
    %203 = arith.constant 127 : i32
    %204 = arith.constant 8 : i32
    %205 = arith.extsi %203 : i32 to i64
    %206 = arith.extsi %204 : i32 to i64
    %207 = llvm.getelementptr %142[%206] : (!llvm.ptr, i64) -> !llvm.ptr, i64
    llvm.store %205, %207 : i64, !llvm.ptr
    %208 = arith.constant 111 : i32
    %209 = arith.constant 9 : i32
    %210 = arith.extsi %208 : i32 to i64
    %211 = arith.extsi %209 : i32 to i64
    %212 = llvm.getelementptr %142[%211] : (!llvm.ptr, i64) -> !llvm.ptr, i64
    llvm.store %210, %212 : i64, !llvm.ptr
    %213 = arith.constant 0 : i32
    %214 = arith.index_cast %213 : i32 to index
    %215 = arith.index_cast %137 : i32 to index
    %217 = arith.constant 1 : index
    %218 = arith.constant -1 : index
    %219 = arith.cmpi sle, %214, %215 : index
    %216 = arith.select %219, %217, %218 : index
    cf.br ^bb24(%214 : index)
    ^bb24(%220: index):
    %221 = arith.cmpi slt, %220, %215 : index
    %222 = arith.cmpi sgt, %220, %215 : index
    %223 = arith.select %219, %221, %222 : i1
    cf.cond_br %223, ^bb25(%220 : index), ^bb26(%220 : index)
    ^bb25(%224: index):
      %225 = arith.constant 1 : i32
      %226 = arith.trunci %225 : i32 to i8
      %227 = arith.index_cast %224 : index to i64
      %228 = llvm.getelementptr %139[%227] : (!llvm.ptr, i64) -> !llvm.ptr, i8
      llvm.store %226, %228 : i8, !llvm.ptr
      %229 = arith.addi %224, %216 : index
      cf.br ^bb24(%229 : index)
    ^bb26(%230: index):
    %231 = arith.constant 0 : i32
    %232 = arith.constant 0 : i32
    %233 = arith.trunci %231 : i32 to i8
    %234 = arith.extsi %232 : i32 to i64
    %235 = llvm.getelementptr %139[%234] : (!llvm.ptr, i64) -> !llvm.ptr, i8
    llvm.store %233, %235 : i8, !llvm.ptr
    %236 = arith.constant 1 : i32
    %237 = arith.extsi %236 : i32 to i64
    %238 = llvm.mlir.constant(1 : i64) : i64
    %239 = llvm.alloca %238 x i64 : (i64) -> !llvm.ptr
    llvm.store %237, %239 : i64, !llvm.ptr
    cf.br ^bb27
    ^bb27:
    %240 = arith.constant 2 : i32
    %241 = llvm.load %239 : !llvm.ptr -> i64
    %243 = arith.extsi %240 : i32 to i64
    %242 = arith.muli %243, %241 : i64
    %244 = llvm.load %239 : !llvm.ptr -> i64
    %245 = arith.muli %242, %244 : i64
    %246 = arith.cmpi slt, %245, %137 : i64
    cf.cond_br %246, ^bb28, ^bb29
    ^bb28:
      %248 = llvm.load %239 : !llvm.ptr -> i64
      %249 = llvm.getelementptr %139[%248] : (!llvm.ptr, i64) -> !llvm.ptr, i8
      %247 = llvm.load %249 : !llvm.ptr -> i8
      %250 = arith.constant 1 : i32
      %252 = arith.extsi %247 : i8 to i32
      %251 = arith.cmpi eq, %252, %250 : i32
      cf.cond_br %251, ^bb30, ^bb31
      ^bb30:
        %253 = arith.constant 3 : i32
        %254 = llvm.load %239 : !llvm.ptr -> i64
        %256 = arith.extsi %253 : i32 to i64
        %255 = arith.muli %256, %254 : i64
        %257 = arith.constant 1 : i32
        %259 = arith.extsi %257 : i32 to i64
        %258 = arith.addi %255, %259 : i64
        %260 = llvm.mlir.constant(1 : i64) : i64
        %261 = llvm.alloca %260 x i64 : (i64) -> !llvm.ptr
        llvm.store %258, %261 : i64, !llvm.ptr
        %262 = arith.constant 2 : i32
        %263 = llvm.load %239 : !llvm.ptr -> i64
        %265 = arith.extsi %262 : i32 to i64
        %264 = arith.muli %265, %263 : i64
        %266 = arith.constant 1 : i32
        %268 = arith.extsi %266 : i32 to i64
        %267 = arith.addi %264, %268 : i64
        cf.br ^bb33
        ^bb33:
        %269 = llvm.load %261 : !llvm.ptr -> i64
        %270 = arith.cmpi slt, %269, %137 : i64
        cf.cond_br %270, ^bb34, ^bb35
        ^bb34:
          %271 = arith.constant 0 : i32
          %272 = llvm.load %261 : !llvm.ptr -> i64
          %273 = arith.trunci %271 : i32 to i8
          %274 = llvm.getelementptr %139[%272] : (!llvm.ptr, i64) -> !llvm.ptr, i8
          llvm.store %273, %274 : i8, !llvm.ptr
          %275 = llvm.load %261 : !llvm.ptr -> i64
          %276 = arith.addi %275, %267 : i64
          llvm.store %276, %261 : i64, !llvm.ptr
          cf.br ^bb33
        ^bb35:
        cf.br ^bb32
      ^bb31:
        cf.br ^bb32
      ^bb32:
      %277 = llvm.load %239 : !llvm.ptr -> i64
      %278 = arith.constant 1 : i32
      %280 = arith.extsi %278 : i32 to i64
      %279 = arith.addi %277, %280 : i64
      llvm.store %279, %239 : i64, !llvm.ptr
      cf.br ^bb27
    ^bb29:
    %281 = arith.constant 0 : i32
    %282 = arith.extsi %281 : i32 to i64
    %283 = llvm.mlir.constant(1 : i64) : i64
    %284 = llvm.alloca %283 x i64 : (i64) -> !llvm.ptr
    llvm.store %282, %284 : i64, !llvm.ptr
    %285 = arith.constant 0 : i32
    %286 = arith.extsi %285 : i32 to i64
    %287 = llvm.mlir.constant(1 : i64) : i64
    %288 = llvm.alloca %287 x i64 : (i64) -> !llvm.ptr
    llvm.store %286, %288 : i64, !llvm.ptr
    %289 = llvm.mlir.constant(1 : i64) : i64
    %290 = llvm.alloca %289 x i64 : (i64) -> !llvm.ptr
    llvm.store %130, %290 : i64, !llvm.ptr
    %291 = llvm.load %290 : !llvm.ptr -> i64
    %292 = arith.constant 2 : i32
    %294 = arith.extsi %292 : i32 to i64
    %293 = arith.remsi %291, %294 : i64
    %295 = arith.constant 0 : i32
    %297 = arith.extsi %295 : i32 to i64
    %296 = arith.cmpi eq, %293, %297 : i64
    cf.cond_br %296, ^bb36, ^bb37
    ^bb36:
      %298 = llvm.load %290 : !llvm.ptr -> i64
      %299 = arith.constant 1 : i32
      %301 = arith.extsi %299 : i32 to i64
      %300 = arith.addi %298, %301 : i64
      llvm.store %300, %290 : i64, !llvm.ptr
      cf.br ^bb38
    ^bb37:
      cf.br ^bb38
    ^bb38:
    cf.br ^bb39
    ^bb39:
    %302 = llvm.load %290 : !llvm.ptr -> i64
    %303 = arith.cmpi sle, %302, %132 : i64
    cf.cond_br %303, ^bb40, ^bb41
    ^bb40:
      %305 = llvm.load %290 : !llvm.ptr -> i64
      %306 = arith.constant 1 : i32
      %308 = arith.extsi %306 : i32 to i64
      %307 = arith.shrsi %305, %308 : i64
      %309 = llvm.getelementptr %139[%307] : (!llvm.ptr, i64) -> !llvm.ptr, i8
      %304 = llvm.load %309 : !llvm.ptr -> i8
      %310 = arith.constant 1 : i32
      %312 = arith.extsi %304 : i8 to i32
      %311 = arith.cmpi eq, %312, %310 : i32
      cf.cond_br %311, ^bb42, ^bb43
      ^bb42:
        %313 = llvm.load %284 : !llvm.ptr -> i64
        %315 = llvm.load %290 : !llvm.ptr -> i64
        %314 = func.call @sam(%315, %142, %147) : (i64, !llvm.ptr, !llvm.ptr) -> i32
        %316 = arith.extsi %314 : i32 to i64
        %317 = arith.addi %313, %316 : i64
        llvm.store %317, %284 : i64, !llvm.ptr
        %318 = llvm.load %288 : !llvm.ptr -> i64
        %320 = llvm.load %290 : !llvm.ptr -> i64
        %321 = arith.constant 0 : i32
        %322 = arith.extsi %321 : i32 to i64
        %319 = func.call @max_clock(%320, %322, %142) : (i64, i64, !llvm.ptr) -> i32
        %323 = arith.extsi %319 : i32 to i64
        %324 = arith.addi %318, %323 : i64
        llvm.store %324, %288 : i64, !llvm.ptr
        cf.br ^bb44
      ^bb43:
        cf.br ^bb44
      ^bb44:
      %325 = llvm.load %290 : !llvm.ptr -> i64
      %326 = arith.constant 2 : i32
      %328 = arith.extsi %326 : i32 to i64
      %327 = arith.addi %325, %328 : i64
      llvm.store %327, %290 : i64, !llvm.ptr
      cf.br ^bb39
    ^bb41:
    %329 = llvm.mlir.addressof @str_0 : !llvm.ptr
    %330 = llvm.load %284 : !llvm.ptr -> i64
    %331 = llvm.load %288 : !llvm.ptr -> i64
    %332 = arith.subi %330, %331 : i64
    %333 = llvm.call @printf(%329, %332) vararg(!llvm.func<i32 (ptr, ...)>) : (!llvm.ptr, i64) -> i32
    func.call @free(%139) : (!llvm.ptr) -> ()
    func.call @free(%142) : (!llvm.ptr) -> ()
    func.call @free(%147) : (!llvm.ptr) -> ()
    %337 = arith.constant 0 : i32
    func.return %337 : i32
  }
}