Problem 377
Last 9 digits of sum_{i=1..17} f(13^i) for digit-sum numbers without 0.
View problem on Project Euler
Performance comparison
| Metric | Our solution | Best known |
|---|
| Time complexity | O(n^3) | O(n) |
|---|
| Space complexity | O(n^2) | O(n) |
|---|
| Approach | Flow solution | Big-integer arithmetic |
|---|
| Verdict | Suboptimal |
|---|
Flow source
# Project Euler 377
# Last 9 digits of sum_{i=1..17} f(13^i) for digit-sum numbers without 0.
extern {
function calloc(n: i64, size: i64) -> ptr<void>
function free(p: ptr<void>) -> void
}
function mat_mul(A: ptr<i64>, B: ptr<i64>, out: ptr<i64>, mod: i64) -> void {
let mut i: i32 = 0
while i < 18 {
let mut j: i32 = 0
while j < 18 {
let mut s: i64 = 0
let mut t: i32 = 0
while t < 18 {
s = (s + A[i * 18 + t] * B[t * 18 + j]) % mod
t = t + 1
}
out[i * 18 + j] = s
j = j + 1
}
i = i + 1
}
}
function mat_vec(A: ptr<i64>, v: ptr<i64>, out: ptr<i64>, mod: i64) -> void {
let mut i: i32 = 0
while i < 18 {
let mut s: i64 = 0
let mut j: i32 = 0
while j < 18 {
s = (s + A[i * 18 + j] * v[j]) % mod
j = j + 1
}
out[i] = s
i = i + 1
}
}
function mat_copy(src: ptr<i64>, dst: ptr<i64>) -> void {
let mut i: i32 = 0
while i < 324 {
dst[i] = src[i]
i = i + 1
}
}
function vec_copy(src: ptr<i64>, dst: ptr<i64>) -> void {
let mut i: i32 = 0
while i < 18 {
dst[i] = src[i]
i = i + 1
}
}
function f_of(n0: i64, powers: ptr<i64>, v0: ptr<i64>, mod: i64) -> i64 {
let v: ptr<i64> = calloc(18, 8)
let tmp: ptr<i64> = calloc(18, 8)
if v == null || tmp == null { return 0 }
vec_copy(v0, v)
let mut n: i64 = n0
let mut bit: i32 = 0
while n > 0 {
if n % 2 == 1 {
mat_vec(powers + (bit as i64) * 324, v, tmp, mod)
vec_copy(tmp, v)
}
n = n / 2
bit = bit + 1
}
let ans: i64 = v[9]
free(tmp)
free(v)
return ans
}
function main() -> i32 {
let MOD: i64 = 1000000000
let powers: ptr<i64> = calloc(64 * 324, 8)
let tmp: ptr<i64> = calloc(324, 8)
let v0: ptr<i64> = calloc(18, 8)
if powers == null || tmp == null || v0 == null { return 1 }
# Build M into powers[0]
let M: ptr<i64> = powers
let mut j: i32 = 0
while j < 9 {
M[0 * 18 + j] = 10 % MOD
j = j + 1
}
let mut i: i32 = 1
while i < 9 {
M[i * 18 + (i - 1)] = 1
i = i + 1
}
j = 0
while j < 9 {
M[9 * 18 + j] = (j + 1) as i64
M[9 * 18 + 9 + j] = 1
j = j + 1
}
i = 10
while i < 18 {
M[i * 18 + (i - 1)] = 1
i = i + 1
}
let mut b: i32 = 1
while b < 64 {
mat_mul(powers + ((b - 1) as i64) * 324, powers + ((b - 1) as i64) * 324, tmp, MOD)
mat_copy(tmp, powers + (b as i64) * 324)
b = b + 1
}
v0[0] = 1
let mut total: i64 = 0
let mut pow13: i64 = 13
let mut ii: i32 = 1
while ii <= 17 {
total = (total + f_of(pow13, powers, v0, MOD)) % MOD
if ii < 17 { pow13 = pow13 * 13 }
ii = ii + 1
}
printf("%lld\n", total)
free(v0)
free(tmp)
free(powers)
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; }
void mat_mul_ptr_i64_ptr_i64_ptr_i64_i64(int64_t* A, int64_t* B, int64_t* out, int64_t mod);
void mat_vec_ptr_i64_ptr_i64_ptr_i64_i64(int64_t* A, int64_t* v, int64_t* out, int64_t mod);
void mat_copy_ptr_i64_ptr_i64(int64_t* src, int64_t* dst);
void vec_copy_ptr_i64_ptr_i64(int64_t* src, int64_t* dst);
int64_t f_of_i64_ptr_i64_ptr_i64_i64(int64_t n0, int64_t* powers, int64_t* v0, int64_t mod);
int32_t main(void);
void mat_mul_ptr_i64_ptr_i64_ptr_i64_i64(int64_t* A, int64_t* B, int64_t* out, int64_t mod) {
int32_t i = 0;
while (i < 18) {
int32_t j = 0;
while (j < 18) {
int64_t s = 0;
int32_t t = 0;
while (t < 18) {
s = FLOW_CHECKED_MOD(((s + (A[((i * 18) + t)] * B[((t * 18) + j)]))), (mod));
t = (t + 1);
}
out[((i * 18) + j)] = s;
j = (j + 1);
}
i = (i + 1);
}
}
void mat_vec_ptr_i64_ptr_i64_ptr_i64_i64(int64_t* A, int64_t* v, int64_t* out, int64_t mod) {
int32_t i = 0;
while (i < 18) {
int64_t s = 0;
int32_t j = 0;
while (j < 18) {
s = FLOW_CHECKED_MOD(((s + (A[((i * 18) + j)] * v[j]))), (mod));
j = (j + 1);
}
out[i] = s;
i = (i + 1);
}
}
void mat_copy_ptr_i64_ptr_i64(int64_t* src, int64_t* dst) {
int32_t i = 0;
while (i < 324) {
dst[i] = src[i];
i = (i + 1);
}
}
void vec_copy_ptr_i64_ptr_i64(int64_t* src, int64_t* dst) {
int32_t i = 0;
while (i < 18) {
dst[i] = src[i];
i = (i + 1);
}
}
int64_t f_of_i64_ptr_i64_ptr_i64_i64(int64_t n0, int64_t* powers, int64_t* v0, int64_t mod) {
int64_t* v = (int64_t*)(calloc(18, 8));
int64_t* tmp = (int64_t*)(calloc(18, 8));
if ((v == NULL || tmp == NULL)) {
return 0;
}
vec_copy_ptr_i64_ptr_i64(v0, v);
int64_t n = n0;
int32_t bit = 0;
while (n > 0) {
if (FLOW_CHECKED_MOD((n), (2)) == 1) {
mat_vec_ptr_i64_ptr_i64_ptr_i64_i64((powers + (((int64_t)(bit)) * 324)), v, tmp, mod);
vec_copy_ptr_i64_ptr_i64(tmp, v);
}
n = FLOW_CHECKED_DIV((n), (2));
bit = (bit + 1);
}
int64_t ans = v[9];
free(tmp);
free(v);
return ans;
}
int32_t main(void) {
int64_t MOD = 1000000000;
int64_t* powers = (int64_t*)(calloc((64 * 324), 8));
int64_t* tmp = (int64_t*)(calloc(324, 8));
int64_t* v0 = (int64_t*)(calloc(18, 8));
if (((powers == NULL || tmp == NULL) || v0 == NULL)) {
return 1;
}
int64_t* M = (int64_t*)(powers);
int32_t j = 0;
while (j < 9) {
M[((0 * 18) + j)] = FLOW_CHECKED_MOD((10), (MOD));
j = (j + 1);
}
int32_t i = 1;
while (i < 9) {
M[((i * 18) + (i - 1))] = 1;
i = (i + 1);
}
j = 0;
while (j < 9) {
M[((9 * 18) + j)] = ((int64_t)((j + 1)));
M[(((9 * 18) + 9) + j)] = 1;
j = (j + 1);
}
i = 10;
while (i < 18) {
M[((i * 18) + (i - 1))] = 1;
i = (i + 1);
}
int32_t b = 1;
while (b < 64) {
mat_mul_ptr_i64_ptr_i64_ptr_i64_i64((powers + (((int64_t)((b - 1))) * 324)), (powers + (((int64_t)((b - 1))) * 324)), tmp, MOD);
mat_copy_ptr_i64_ptr_i64(tmp, (powers + (((int64_t)(b)) * 324)));
b = (b + 1);
}
v0[0] = 1;
int64_t total = 0;
int64_t pow13 = 13;
int32_t ii = 1;
while (ii <= 17) {
total = FLOW_CHECKED_MOD(((total + f_of_i64_ptr_i64_ptr_i64_i64(pow13, powers, v0, MOD))), (MOD));
if (ii < 17) {
pow13 = (pow13 * 13);
}
ii = (ii + 1);
}
printf("%lld\n", total);
free(v0);
free(tmp);
free(powers);
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 @mat_mul(%arg0: !llvm.ptr, %arg1: !llvm.ptr, %arg2: !llvm.ptr, %arg3: i64) -> () {
%0 = arith.constant 0 : i32
%1 = llvm.mlir.constant(1 : i64) : i64
%2 = llvm.alloca %1 x i32 : (i64) -> !llvm.ptr
llvm.store %0, %2 : i32, !llvm.ptr
cf.br ^bb0
^bb0:
%3 = llvm.load %2 : !llvm.ptr -> i32
%4 = arith.constant 18 : i32
%5 = arith.cmpi slt, %3, %4 : i32
cf.cond_br %5, ^bb1, ^bb2
^bb1:
%6 = arith.constant 0 : i32
%7 = llvm.mlir.constant(1 : i64) : i64
%8 = llvm.alloca %7 x i32 : (i64) -> !llvm.ptr
llvm.store %6, %8 : i32, !llvm.ptr
cf.br ^bb3
^bb3:
%9 = llvm.load %8 : !llvm.ptr -> i32
%10 = arith.constant 18 : i32
%11 = arith.cmpi slt, %9, %10 : i32
cf.cond_br %11, ^bb4, ^bb5
^bb4:
%12 = arith.constant 0 : i32
%13 = arith.extsi %12 : i32 to i64
%14 = llvm.mlir.constant(1 : i64) : i64
%15 = llvm.alloca %14 x i64 : (i64) -> !llvm.ptr
llvm.store %13, %15 : i64, !llvm.ptr
%16 = arith.constant 0 : i32
%17 = llvm.mlir.constant(1 : i64) : i64
%18 = llvm.alloca %17 x i32 : (i64) -> !llvm.ptr
llvm.store %16, %18 : i32, !llvm.ptr
cf.br ^bb6
^bb6:
%19 = llvm.load %18 : !llvm.ptr -> i32
%20 = arith.constant 18 : i32
%21 = arith.cmpi slt, %19, %20 : i32
cf.cond_br %21, ^bb7, ^bb8
^bb7:
%22 = llvm.load %15 : !llvm.ptr -> i64
%24 = llvm.load %2 : !llvm.ptr -> i32
%25 = arith.constant 18 : i32
%26 = arith.muli %24, %25 : i32
%27 = llvm.load %18 : !llvm.ptr -> i32
%28 = arith.addi %26, %27 : i32
%29 = arith.extsi %28 : i32 to i64
%30 = llvm.getelementptr %arg0[%29] : (!llvm.ptr, i64) -> !llvm.ptr, i64
%23 = llvm.load %30 : !llvm.ptr -> i64
%32 = llvm.load %18 : !llvm.ptr -> i32
%33 = arith.constant 18 : i32
%34 = arith.muli %32, %33 : i32
%35 = llvm.load %8 : !llvm.ptr -> i32
%36 = arith.addi %34, %35 : i32
%37 = arith.extsi %36 : i32 to i64
%38 = llvm.getelementptr %arg1[%37] : (!llvm.ptr, i64) -> !llvm.ptr, i64
%31 = llvm.load %38 : !llvm.ptr -> i64
%39 = arith.muli %23, %31 : i64
%40 = arith.addi %22, %39 : i64
%41 = arith.remsi %40, %arg3 : i64
llvm.store %41, %15 : i64, !llvm.ptr
%42 = llvm.load %18 : !llvm.ptr -> i32
%43 = arith.constant 1 : i32
%44 = arith.addi %42, %43 : i32
llvm.store %44, %18 : i32, !llvm.ptr
cf.br ^bb6
^bb8:
%45 = llvm.load %15 : !llvm.ptr -> i64
%46 = llvm.load %2 : !llvm.ptr -> i32
%47 = arith.constant 18 : i32
%48 = arith.muli %46, %47 : i32
%49 = llvm.load %8 : !llvm.ptr -> i32
%50 = arith.addi %48, %49 : i32
%51 = arith.extsi %50 : i32 to i64
%52 = llvm.getelementptr %arg2[%51] : (!llvm.ptr, i64) -> !llvm.ptr, i64
llvm.store %45, %52 : i64, !llvm.ptr
%53 = llvm.load %8 : !llvm.ptr -> i32
%54 = arith.constant 1 : i32
%55 = arith.addi %53, %54 : i32
llvm.store %55, %8 : i32, !llvm.ptr
cf.br ^bb3
^bb5:
%56 = llvm.load %2 : !llvm.ptr -> i32
%57 = arith.constant 1 : i32
%58 = arith.addi %56, %57 : i32
llvm.store %58, %2 : i32, !llvm.ptr
cf.br ^bb0
^bb2:
func.return
}
func.func @mat_vec(%arg0: !llvm.ptr, %arg1: !llvm.ptr, %arg2: !llvm.ptr, %arg3: i64) -> () {
%59 = arith.constant 0 : i32
%60 = llvm.mlir.constant(1 : i64) : i64
%61 = llvm.alloca %60 x i32 : (i64) -> !llvm.ptr
llvm.store %59, %61 : i32, !llvm.ptr
cf.br ^bb9
^bb9:
%62 = llvm.load %61 : !llvm.ptr -> i32
%63 = arith.constant 18 : i32
%64 = arith.cmpi slt, %62, %63 : i32
cf.cond_br %64, ^bb10, ^bb11
^bb10:
%65 = arith.constant 0 : i32
%66 = arith.extsi %65 : i32 to i64
%67 = llvm.mlir.constant(1 : i64) : i64
%68 = llvm.alloca %67 x i64 : (i64) -> !llvm.ptr
llvm.store %66, %68 : i64, !llvm.ptr
%69 = arith.constant 0 : i32
%70 = llvm.mlir.constant(1 : i64) : i64
%71 = llvm.alloca %70 x i32 : (i64) -> !llvm.ptr
llvm.store %69, %71 : i32, !llvm.ptr
cf.br ^bb12
^bb12:
%72 = llvm.load %71 : !llvm.ptr -> i32
%73 = arith.constant 18 : i32
%74 = arith.cmpi slt, %72, %73 : i32
cf.cond_br %74, ^bb13, ^bb14
^bb13:
%75 = llvm.load %68 : !llvm.ptr -> i64
%77 = llvm.load %61 : !llvm.ptr -> i32
%78 = arith.constant 18 : i32
%79 = arith.muli %77, %78 : i32
%80 = llvm.load %71 : !llvm.ptr -> i32
%81 = arith.addi %79, %80 : i32
%82 = arith.extsi %81 : i32 to i64
%83 = llvm.getelementptr %arg0[%82] : (!llvm.ptr, i64) -> !llvm.ptr, i64
%76 = llvm.load %83 : !llvm.ptr -> i64
%85 = llvm.load %71 : !llvm.ptr -> i32
%86 = arith.extsi %85 : i32 to i64
%87 = llvm.getelementptr %arg1[%86] : (!llvm.ptr, i64) -> !llvm.ptr, i64
%84 = llvm.load %87 : !llvm.ptr -> i64
%88 = arith.muli %76, %84 : i64
%89 = arith.addi %75, %88 : i64
%90 = arith.remsi %89, %arg3 : i64
llvm.store %90, %68 : i64, !llvm.ptr
%91 = llvm.load %71 : !llvm.ptr -> i32
%92 = arith.constant 1 : i32
%93 = arith.addi %91, %92 : i32
llvm.store %93, %71 : i32, !llvm.ptr
cf.br ^bb12
^bb14:
%94 = llvm.load %68 : !llvm.ptr -> i64
%95 = llvm.load %61 : !llvm.ptr -> i32
%96 = arith.extsi %95 : i32 to i64
%97 = llvm.getelementptr %arg2[%96] : (!llvm.ptr, i64) -> !llvm.ptr, i64
llvm.store %94, %97 : i64, !llvm.ptr
%98 = llvm.load %61 : !llvm.ptr -> i32
%99 = arith.constant 1 : i32
%100 = arith.addi %98, %99 : i32
llvm.store %100, %61 : i32, !llvm.ptr
cf.br ^bb9
^bb11:
func.return
}
func.func @mat_copy(%arg0: !llvm.ptr, %arg1: !llvm.ptr) -> () {
%101 = arith.constant 0 : i32
%102 = llvm.mlir.constant(1 : i64) : i64
%103 = llvm.alloca %102 x i32 : (i64) -> !llvm.ptr
llvm.store %101, %103 : i32, !llvm.ptr
cf.br ^bb15
^bb15:
%104 = llvm.load %103 : !llvm.ptr -> i32
%105 = arith.constant 324 : i32
%106 = arith.cmpi slt, %104, %105 : i32
cf.cond_br %106, ^bb16, ^bb17
^bb16:
%108 = llvm.load %103 : !llvm.ptr -> i32
%109 = arith.extsi %108 : i32 to i64
%110 = llvm.getelementptr %arg0[%109] : (!llvm.ptr, i64) -> !llvm.ptr, i64
%107 = llvm.load %110 : !llvm.ptr -> i64
%111 = llvm.load %103 : !llvm.ptr -> i32
%112 = arith.extsi %111 : i32 to i64
%113 = llvm.getelementptr %arg1[%112] : (!llvm.ptr, i64) -> !llvm.ptr, i64
llvm.store %107, %113 : i64, !llvm.ptr
%114 = llvm.load %103 : !llvm.ptr -> i32
%115 = arith.constant 1 : i32
%116 = arith.addi %114, %115 : i32
llvm.store %116, %103 : i32, !llvm.ptr
cf.br ^bb15
^bb17:
func.return
}
func.func @vec_copy(%arg0: !llvm.ptr, %arg1: !llvm.ptr) -> () {
%117 = arith.constant 0 : i32
%118 = llvm.mlir.constant(1 : i64) : i64
%119 = llvm.alloca %118 x i32 : (i64) -> !llvm.ptr
llvm.store %117, %119 : i32, !llvm.ptr
cf.br ^bb18
^bb18:
%120 = llvm.load %119 : !llvm.ptr -> i32
%121 = arith.constant 18 : i32
%122 = arith.cmpi slt, %120, %121 : i32
cf.cond_br %122, ^bb19, ^bb20
^bb19:
%124 = llvm.load %119 : !llvm.ptr -> i32
%125 = arith.extsi %124 : i32 to i64
%126 = llvm.getelementptr %arg0[%125] : (!llvm.ptr, i64) -> !llvm.ptr, i64
%123 = llvm.load %126 : !llvm.ptr -> i64
%127 = llvm.load %119 : !llvm.ptr -> i32
%128 = arith.extsi %127 : i32 to i64
%129 = llvm.getelementptr %arg1[%128] : (!llvm.ptr, i64) -> !llvm.ptr, i64
llvm.store %123, %129 : i64, !llvm.ptr
%130 = llvm.load %119 : !llvm.ptr -> i32
%131 = arith.constant 1 : i32
%132 = arith.addi %130, %131 : i32
llvm.store %132, %119 : i32, !llvm.ptr
cf.br ^bb18
^bb20:
func.return
}
func.func @f_of(%arg0: i64, %arg1: !llvm.ptr, %arg2: !llvm.ptr, %arg3: i64) -> i64 {
%134 = arith.constant 18 : i32
%135 = arith.constant 8 : i32
%136 = arith.extsi %134 : i32 to i64
%137 = arith.extsi %135 : i32 to i64
%133 = func.call @calloc(%136, %137) : (i64, i64) -> !llvm.ptr
%139 = arith.constant 18 : i32
%140 = arith.constant 8 : i32
%141 = arith.extsi %139 : i32 to i64
%142 = arith.extsi %140 : i32 to i64
%138 = func.call @calloc(%141, %142) : (i64, i64) -> !llvm.ptr
%143 = llvm.mlir.zero : !llvm.ptr
%144 = llvm.icmp "eq" %133, %143 : !llvm.ptr
%145 = scf.if %144 -> (i1) {
%146 = arith.constant true
scf.yield %146 : i1
} else {
%147 = llvm.mlir.zero : !llvm.ptr
%148 = llvm.icmp "eq" %138, %147 : !llvm.ptr
scf.yield %148 : i1
}
cf.cond_br %145, ^bb21, ^bb22
^bb21:
%149 = arith.constant 0 : i32
%150 = arith.extsi %149 : i32 to i64
func.return %150 : i64
^bb22:
cf.br ^bb23
^bb23:
func.call @vec_copy(%arg2, %133) : (!llvm.ptr, !llvm.ptr) -> ()
%152 = llvm.mlir.constant(1 : i64) : i64
%153 = llvm.alloca %152 x i64 : (i64) -> !llvm.ptr
llvm.store %arg0, %153 : i64, !llvm.ptr
%154 = arith.constant 0 : i32
%155 = llvm.mlir.constant(1 : i64) : i64
%156 = llvm.alloca %155 x i32 : (i64) -> !llvm.ptr
llvm.store %154, %156 : i32, !llvm.ptr
cf.br ^bb24
^bb24:
%157 = llvm.load %153 : !llvm.ptr -> i64
%158 = arith.constant 0 : i32
%160 = arith.extsi %158 : i32 to i64
%159 = arith.cmpi sgt, %157, %160 : i64
cf.cond_br %159, ^bb25, ^bb26
^bb25:
%161 = llvm.load %153 : !llvm.ptr -> i64
%162 = arith.constant 2 : i32
%164 = arith.extsi %162 : i32 to i64
%163 = arith.remsi %161, %164 : i64
%165 = arith.constant 1 : i32
%167 = arith.extsi %165 : i32 to i64
%166 = arith.cmpi eq, %163, %167 : i64
cf.cond_br %166, ^bb27, ^bb28
^bb27:
# String concatenation: !llvm.ptr + i64
func.call @mat_vec(%169, %133, %138, %arg3) : (!llvm.ptr, !llvm.ptr, !llvm.ptr, i64) -> ()
func.call @vec_copy(%138, %133) : (!llvm.ptr, !llvm.ptr) -> ()
cf.br ^bb29
^bb28:
cf.br ^bb29
^bb29:
%171 = llvm.load %153 : !llvm.ptr -> i64
%172 = arith.constant 2 : i32
%174 = arith.extsi %172 : i32 to i64
%173 = arith.divsi %171, %174 : i64
llvm.store %173, %153 : i64, !llvm.ptr
%175 = llvm.load %156 : !llvm.ptr -> i32
%176 = arith.constant 1 : i32
%177 = arith.addi %175, %176 : i32
llvm.store %177, %156 : i32, !llvm.ptr
cf.br ^bb24
^bb26:
%179 = arith.constant 9 : i32
%180 = arith.extsi %179 : i32 to i64
%181 = llvm.getelementptr %133[%180] : (!llvm.ptr, i64) -> !llvm.ptr, i64
%178 = llvm.load %181 : !llvm.ptr -> i64
func.call @free(%138) : (!llvm.ptr) -> ()
func.call @free(%133) : (!llvm.ptr) -> ()
func.return %178 : i64
}
func.func @main() -> i32 {
%184 = arith.constant 1000000000 : i32
%185 = arith.extsi %184 : i32 to i64
%187 = arith.constant 64 : i32
%188 = arith.constant 324 : i32
%189 = arith.muli %187, %188 : i32
%190 = arith.constant 8 : i32
%191 = arith.extsi %189 : i32 to i64
%192 = arith.extsi %190 : i32 to i64
%186 = func.call @calloc(%191, %192) : (i64, i64) -> !llvm.ptr
%194 = arith.constant 324 : i32
%195 = arith.constant 8 : i32
%196 = arith.extsi %194 : i32 to i64
%197 = arith.extsi %195 : i32 to i64
%193 = func.call @calloc(%196, %197) : (i64, i64) -> !llvm.ptr
%199 = arith.constant 18 : i32
%200 = arith.constant 8 : i32
%201 = arith.extsi %199 : i32 to i64
%202 = arith.extsi %200 : i32 to i64
%198 = func.call @calloc(%201, %202) : (i64, i64) -> !llvm.ptr
%203 = llvm.mlir.zero : !llvm.ptr
%204 = llvm.icmp "eq" %186, %203 : !llvm.ptr
%205 = scf.if %204 -> (i1) {
%206 = arith.constant true
scf.yield %206 : i1
} else {
%207 = llvm.mlir.zero : !llvm.ptr
%208 = llvm.icmp "eq" %193, %207 : !llvm.ptr
scf.yield %208 : i1
}
%209 = scf.if %205 -> (i1) {
%210 = arith.constant true
scf.yield %210 : i1
} else {
%211 = llvm.mlir.zero : !llvm.ptr
%212 = llvm.icmp "eq" %198, %211 : !llvm.ptr
scf.yield %212 : i1
}
cf.cond_br %209, ^bb30, ^bb31
^bb30:
%213 = arith.constant 1 : i32
func.return %213 : i32
^bb31:
cf.br ^bb32
^bb32:
%214 = arith.constant 0 : i32
%215 = llvm.mlir.constant(1 : i64) : i64
%216 = llvm.alloca %215 x i32 : (i64) -> !llvm.ptr
llvm.store %214, %216 : i32, !llvm.ptr
cf.br ^bb33
^bb33:
%217 = llvm.load %216 : !llvm.ptr -> i32
%218 = arith.constant 9 : i32
%219 = arith.cmpi slt, %217, %218 : i32
cf.cond_br %219, ^bb34, ^bb35
^bb34:
%220 = arith.constant 10 : i32
%222 = arith.extsi %220 : i32 to i64
%221 = arith.remsi %222, %185 : i64
%223 = arith.constant 0 : i32
%224 = arith.constant 18 : i32
%225 = arith.muli %223, %224 : i32
%226 = llvm.load %216 : !llvm.ptr -> i32
%227 = arith.addi %225, %226 : i32
%228 = arith.extsi %227 : i32 to i64
%229 = llvm.getelementptr %186[%228] : (!llvm.ptr, i64) -> !llvm.ptr, i64
llvm.store %221, %229 : i64, !llvm.ptr
%230 = llvm.load %216 : !llvm.ptr -> i32
%231 = arith.constant 1 : i32
%232 = arith.addi %230, %231 : i32
llvm.store %232, %216 : i32, !llvm.ptr
cf.br ^bb33
^bb35:
%233 = arith.constant 1 : i32
%234 = llvm.mlir.constant(1 : i64) : i64
%235 = llvm.alloca %234 x i32 : (i64) -> !llvm.ptr
llvm.store %233, %235 : i32, !llvm.ptr
cf.br ^bb36
^bb36:
%236 = llvm.load %235 : !llvm.ptr -> i32
%237 = arith.constant 9 : i32
%238 = arith.cmpi slt, %236, %237 : i32
cf.cond_br %238, ^bb37, ^bb38
^bb37:
%239 = arith.constant 1 : i32
%240 = llvm.load %235 : !llvm.ptr -> i32
%241 = arith.constant 18 : i32
%242 = arith.muli %240, %241 : i32
%243 = llvm.load %235 : !llvm.ptr -> i32
%244 = arith.constant 1 : i32
%245 = arith.subi %243, %244 : i32
%246 = arith.addi %242, %245 : i32
%247 = arith.extsi %239 : i32 to i64
%248 = arith.extsi %246 : i32 to i64
%249 = llvm.getelementptr %186[%248] : (!llvm.ptr, i64) -> !llvm.ptr, i64
llvm.store %247, %249 : i64, !llvm.ptr
%250 = llvm.load %235 : !llvm.ptr -> i32
%251 = arith.constant 1 : i32
%252 = arith.addi %250, %251 : i32
llvm.store %252, %235 : i32, !llvm.ptr
cf.br ^bb36
^bb38:
%253 = arith.constant 0 : i32
llvm.store %253, %216 : i32, !llvm.ptr
cf.br ^bb39
^bb39:
%254 = llvm.load %216 : !llvm.ptr -> i32
%255 = arith.constant 9 : i32
%256 = arith.cmpi slt, %254, %255 : i32
cf.cond_br %256, ^bb40, ^bb41
^bb40:
%257 = llvm.load %216 : !llvm.ptr -> i32
%258 = arith.constant 1 : i32
%259 = arith.addi %257, %258 : i32
%260 = arith.extsi %259 : i32 to i64
%261 = arith.constant 9 : i32
%262 = arith.constant 18 : i32
%263 = arith.muli %261, %262 : i32
%264 = llvm.load %216 : !llvm.ptr -> i32
%265 = arith.addi %263, %264 : i32
%266 = arith.extsi %265 : i32 to i64
%267 = llvm.getelementptr %186[%266] : (!llvm.ptr, i64) -> !llvm.ptr, i64
llvm.store %260, %267 : i64, !llvm.ptr
%268 = arith.constant 1 : i32
%269 = arith.constant 9 : i32
%270 = arith.constant 18 : i32
%271 = arith.muli %269, %270 : i32
%272 = arith.constant 9 : i32
%273 = arith.addi %271, %272 : i32
%274 = llvm.load %216 : !llvm.ptr -> i32
%275 = arith.addi %273, %274 : i32
%276 = arith.extsi %268 : i32 to i64
%277 = arith.extsi %275 : i32 to i64
%278 = llvm.getelementptr %186[%277] : (!llvm.ptr, i64) -> !llvm.ptr, i64
llvm.store %276, %278 : i64, !llvm.ptr
%279 = llvm.load %216 : !llvm.ptr -> i32
%280 = arith.constant 1 : i32
%281 = arith.addi %279, %280 : i32
llvm.store %281, %216 : i32, !llvm.ptr
cf.br ^bb39
^bb41:
%282 = arith.constant 10 : i32
llvm.store %282, %235 : i32, !llvm.ptr
cf.br ^bb42
^bb42:
%283 = llvm.load %235 : !llvm.ptr -> i32
%284 = arith.constant 18 : i32
%285 = arith.cmpi slt, %283, %284 : i32
cf.cond_br %285, ^bb43, ^bb44
^bb43:
%286 = arith.constant 1 : i32
%287 = llvm.load %235 : !llvm.ptr -> i32
%288 = arith.constant 18 : i32
%289 = arith.muli %287, %288 : i32
%290 = llvm.load %235 : !llvm.ptr -> i32
%291 = arith.constant 1 : i32
%292 = arith.subi %290, %291 : i32
%293 = arith.addi %289, %292 : i32
%294 = arith.extsi %286 : i32 to i64
%295 = arith.extsi %293 : i32 to i64
%296 = llvm.getelementptr %186[%295] : (!llvm.ptr, i64) -> !llvm.ptr, i64
llvm.store %294, %296 : i64, !llvm.ptr
%297 = llvm.load %235 : !llvm.ptr -> i32
%298 = arith.constant 1 : i32
%299 = arith.addi %297, %298 : i32
llvm.store %299, %235 : i32, !llvm.ptr
cf.br ^bb42
^bb44:
%300 = arith.constant 1 : i32
%301 = llvm.mlir.constant(1 : i64) : i64
%302 = llvm.alloca %301 x i32 : (i64) -> !llvm.ptr
llvm.store %300, %302 : i32, !llvm.ptr
cf.br ^bb45
^bb45:
%303 = llvm.load %302 : !llvm.ptr -> i32
%304 = arith.constant 64 : i32
%305 = arith.cmpi slt, %303, %304 : i32
cf.cond_br %305, ^bb46, ^bb47
^bb46:
# String concatenation: !llvm.ptr + i64
# String concatenation: !llvm.ptr + i64
func.call @mat_mul(%307, %308, %193, %185) : (!llvm.ptr, !llvm.ptr, !llvm.ptr, i64) -> ()
# String concatenation: !llvm.ptr + i64
func.call @mat_copy(%193, %310) : (!llvm.ptr, !llvm.ptr) -> ()
%311 = llvm.load %302 : !llvm.ptr -> i32
%312 = arith.constant 1 : i32
%313 = arith.addi %311, %312 : i32
llvm.store %313, %302 : i32, !llvm.ptr
cf.br ^bb45
^bb47:
%314 = arith.constant 1 : i32
%315 = arith.constant 0 : i32
%316 = arith.extsi %314 : i32 to i64
%317 = arith.extsi %315 : i32 to i64
%318 = llvm.getelementptr %198[%317] : (!llvm.ptr, i64) -> !llvm.ptr, i64
llvm.store %316, %318 : i64, !llvm.ptr
%319 = arith.constant 0 : i32
%320 = arith.extsi %319 : i32 to i64
%321 = llvm.mlir.constant(1 : i64) : i64
%322 = llvm.alloca %321 x i64 : (i64) -> !llvm.ptr
llvm.store %320, %322 : i64, !llvm.ptr
%323 = arith.constant 13 : i32
%324 = arith.extsi %323 : i32 to i64
%325 = llvm.mlir.constant(1 : i64) : i64
%326 = llvm.alloca %325 x i64 : (i64) -> !llvm.ptr
llvm.store %324, %326 : i64, !llvm.ptr
%327 = arith.constant 1 : i32
%328 = llvm.mlir.constant(1 : i64) : i64
%329 = llvm.alloca %328 x i32 : (i64) -> !llvm.ptr
llvm.store %327, %329 : i32, !llvm.ptr
cf.br ^bb48
^bb48:
%330 = llvm.load %329 : !llvm.ptr -> i32
%331 = arith.constant 17 : i32
%332 = arith.cmpi sle, %330, %331 : i32
cf.cond_br %332, ^bb49, ^bb50
^bb49:
%333 = llvm.load %322 : !llvm.ptr -> i64
%335 = llvm.load %326 : !llvm.ptr -> i64
%334 = func.call @f_of(%335, %186, %198, %185) : (i64, !llvm.ptr, !llvm.ptr, i64) -> i64
%336 = arith.addi %333, %334 : i64
%337 = arith.remsi %336, %185 : i64
llvm.store %337, %322 : i64, !llvm.ptr
%338 = llvm.load %329 : !llvm.ptr -> i32
%339 = arith.constant 17 : i32
%340 = arith.cmpi slt, %338, %339 : i32
cf.cond_br %340, ^bb51, ^bb52
^bb51:
%341 = llvm.load %326 : !llvm.ptr -> i64
%342 = arith.constant 13 : i32
%344 = arith.extsi %342 : i32 to i64
%343 = arith.muli %341, %344 : i64
llvm.store %343, %326 : i64, !llvm.ptr
cf.br ^bb53
^bb52:
cf.br ^bb53
^bb53:
%345 = llvm.load %329 : !llvm.ptr -> i32
%346 = arith.constant 1 : i32
%347 = arith.addi %345, %346 : i32
llvm.store %347, %329 : i32, !llvm.ptr
cf.br ^bb48
^bb50:
%348 = llvm.mlir.addressof @str_0 : !llvm.ptr
%349 = llvm.load %322 : !llvm.ptr -> i64
%350 = llvm.call @printf(%348, %349) vararg(!llvm.func<i32 (ptr, ...)>) : (!llvm.ptr, i64) -> i32
func.call @free(%198) : (!llvm.ptr) -> ()
func.call @free(%193) : (!llvm.ptr) -> ()
func.call @free(%186) : (!llvm.ptr) -> ()
%354 = arith.constant 0 : i32
func.return %354 : i32
}
}