Problem 132

Sum of first forty prime factors of R(10^9).

View problem on Project Euler

Performance comparison

Flow source

# Project Euler 132
# Sum of first forty prime factors of R(10^9).

function mod_pow(base: i64, exp: i64, mod: i64) -> i64 {
    let mut r: i64 = 1
    let mut b: i64 = base % mod
    let mut e: i64 = exp
    while e > 0 {
        if e % 2 == 1 { r = (r * b) % mod }
        b = (b * b) % mod
        e = e / 2
    }
    return r
}

function main() -> i32 {
    let digits: i64 = 1000000000
    let mut total: i64 = 0
    let mut left: i32 = 40
    let mut n: i64 = 7
    while left > 0 {
        let mut prime: bool = true
        let mut d: i64 = 3
        while d * d <= n {
            if n % d == 0 { prime = false; break }
            d = d + 2
        }
        if prime {
            if mod_pow(10, digits, 9 * n) == 1 {
                total = total + n
                left = left - 1
            }
        }
        n = n + 2
    }
    printf("%lld\n", total)
    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 base, int64_t exp, int64_t mod);
int32_t main(void);

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

int32_t main(void) {
    int64_t digits = 1000000000;
    int64_t total = 0;
    int32_t left = 40;
    int64_t n = 7;
    while (left > 0) {
        bool prime = 1;
        int64_t d = 3;
        while ((d * d) <= n) {
            if (FLOW_CHECKED_MOD((n), (d)) == 0) {
                prime = 0;
                break;
            }
            d = (d + 2);
        }
        if (prime) {
            if (mod_pow_i64_i64_i64(10, digits, (9 * n)) == 1) {
                total = (total + n);
                left = (left - 1);
            }
        }
        n = (n + 2);
    }
    printf("%lld\n", total);
    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 @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 2 : i32
      %16 = arith.extsi %14 : i32 to i64
      %15 = arith.remsi %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 = llvm.load %6 : !llvm.ptr -> i64
        %22 = arith.muli %20, %21 : i64
        %23 = arith.remsi %22, %arg2 : i64
        llvm.store %23, %3 : i64, !llvm.ptr
        cf.br ^bb5
      ^bb4:
        cf.br ^bb5
      ^bb5:
      %24 = llvm.load %6 : !llvm.ptr -> i64
      %25 = llvm.load %6 : !llvm.ptr -> i64
      %26 = arith.muli %24, %25 : i64
      %27 = arith.remsi %26, %arg2 : i64
      llvm.store %27, %6 : i64, !llvm.ptr
      %28 = llvm.load %8 : !llvm.ptr -> i64
      %29 = arith.constant 2 : i32
      %31 = arith.extsi %29 : i32 to i64
      %30 = arith.divsi %28, %31 : i64
      llvm.store %30, %8 : i64, !llvm.ptr
      cf.br ^bb0
    ^bb2:
    %32 = llvm.load %3 : !llvm.ptr -> i64
    func.return %32 : i64
  }
  func.func @main() -> i32 {
    %33 = arith.constant 1000000000 : i32
    %34 = arith.extsi %33 : i32 to i64
    %35 = arith.constant 0 : i32
    %36 = arith.extsi %35 : i32 to i64
    %37 = llvm.mlir.constant(1 : i64) : i64
    %38 = llvm.alloca %37 x i64 : (i64) -> !llvm.ptr
    llvm.store %36, %38 : i64, !llvm.ptr
    %39 = arith.constant 40 : i32
    %40 = llvm.mlir.constant(1 : i64) : i64
    %41 = llvm.alloca %40 x i32 : (i64) -> !llvm.ptr
    llvm.store %39, %41 : i32, !llvm.ptr
    %42 = arith.constant 7 : 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
    cf.br ^bb6
    ^bb6:
    %46 = llvm.load %41 : !llvm.ptr -> i32
    %47 = arith.constant 0 : i32
    %48 = arith.cmpi sgt, %46, %47 : i32
    cf.cond_br %48, ^bb7, ^bb8
    ^bb7:
      %49 = arith.constant 1 : i1
      %50 = llvm.mlir.constant(1 : i64) : i64
      %51 = llvm.alloca %50 x i1 : (i64) -> !llvm.ptr
      llvm.store %49, %51 : i1, !llvm.ptr
      %52 = arith.constant 3 : i32
      %53 = arith.extsi %52 : i32 to i64
      %54 = llvm.mlir.constant(1 : i64) : i64
      %55 = llvm.alloca %54 x i64 : (i64) -> !llvm.ptr
      llvm.store %53, %55 : i64, !llvm.ptr
      cf.br ^bb9
      ^bb9:
      %56 = llvm.load %55 : !llvm.ptr -> i64
      %57 = llvm.load %55 : !llvm.ptr -> i64
      %58 = arith.muli %56, %57 : i64
      %59 = llvm.load %45 : !llvm.ptr -> i64
      %60 = arith.cmpi sle, %58, %59 : i64
      cf.cond_br %60, ^bb10, ^bb11
      ^bb10:
        %61 = llvm.load %45 : !llvm.ptr -> i64
        %62 = llvm.load %55 : !llvm.ptr -> i64
        %63 = arith.remsi %61, %62 : i64
        %64 = arith.constant 0 : i32
        %66 = arith.extsi %64 : i32 to i64
        %65 = arith.cmpi eq, %63, %66 : i64
        cf.cond_br %65, ^bb12, ^bb13
        ^bb12:
          %67 = arith.constant 0 : i1
          llvm.store %67, %51 : i1, !llvm.ptr
          cf.br ^bb11
        ^bb13:
          cf.br ^bb14
        ^bb14:
        %68 = llvm.load %55 : !llvm.ptr -> i64
        %69 = arith.constant 2 : i32
        %71 = arith.extsi %69 : i32 to i64
        %70 = arith.addi %68, %71 : i64
        llvm.store %70, %55 : i64, !llvm.ptr
        cf.br ^bb9
      ^bb11:
      %72 = llvm.load %51 : !llvm.ptr -> i1
      cf.cond_br %72, ^bb15, ^bb16
      ^bb15:
        %74 = arith.constant 10 : i32
        %75 = arith.constant 9 : i32
        %76 = llvm.load %45 : !llvm.ptr -> i64
        %78 = arith.extsi %75 : i32 to i64
        %77 = arith.muli %78, %76 : i64
        %79 = arith.extsi %74 : i32 to i64
        %73 = func.call @mod_pow(%79, %34, %77) : (i64, i64, i64) -> i64
        %80 = arith.constant 1 : i32
        %82 = arith.extsi %80 : i32 to i64
        %81 = arith.cmpi eq, %73, %82 : i64
        cf.cond_br %81, ^bb18, ^bb19
        ^bb18:
          %83 = llvm.load %38 : !llvm.ptr -> i64
          %84 = llvm.load %45 : !llvm.ptr -> i64
          %85 = arith.addi %83, %84 : i64
          llvm.store %85, %38 : i64, !llvm.ptr
          %86 = llvm.load %41 : !llvm.ptr -> i32
          %87 = arith.constant 1 : i32
          %88 = arith.subi %86, %87 : i32
          llvm.store %88, %41 : i32, !llvm.ptr
          cf.br ^bb20
        ^bb19:
          cf.br ^bb20
        ^bb20:
        cf.br ^bb17
      ^bb16:
        cf.br ^bb17
      ^bb17:
      %89 = llvm.load %45 : !llvm.ptr -> i64
      %90 = arith.constant 2 : i32
      %92 = arith.extsi %90 : i32 to i64
      %91 = arith.addi %89, %92 : i64
      llvm.store %91, %45 : i64, !llvm.ptr
      cf.br ^bb6
    ^bb8:
    %93 = llvm.mlir.addressof @str_0 : !llvm.ptr
    %94 = llvm.load %38 : !llvm.ptr -> i64
    %95 = llvm.call @printf(%93, %94) vararg(!llvm.func<i32 (ptr, ...)>) : (!llvm.ptr, i64) -> i32
    %96 = arith.constant 0 : i32
    func.return %96 : i32
  }
}