147 lines
4.5 KiB
C++
147 lines
4.5 KiB
C++
|
/*
|
||
|
* ezSAT -- A simple and easy to use CNF generator for SAT solvers
|
||
|
*
|
||
|
* Copyright (C) 2013 Clifford Wolf <clifford@clifford.at>
|
||
|
*
|
||
|
* Permission to use, copy, modify, and/or distribute this software for any
|
||
|
* purpose with or without fee is hereby granted, provided that the above
|
||
|
* copyright notice and this permission notice appear in all copies.
|
||
|
*
|
||
|
* THE SOFTWARE IS PROVIDED "AS IS" AND THE AUTHOR DISCLAIMS ALL WARRANTIES
|
||
|
* WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES OF
|
||
|
* MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR
|
||
|
* ANY SPECIAL, DIRECT, INDIRECT, OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES
|
||
|
* WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN AN
|
||
|
* ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT OF
|
||
|
* OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE.
|
||
|
*
|
||
|
*/
|
||
|
|
||
|
#include "ezminisat.h"
|
||
|
#include <assert.h>
|
||
|
|
||
|
#define INIT_X 123456789
|
||
|
#define INIT_Y 362436069
|
||
|
#define INIT_Z 521288629
|
||
|
#define INIT_W 88675123
|
||
|
|
||
|
uint32_t xorshift128() {
|
||
|
static uint32_t x = INIT_X;
|
||
|
static uint32_t y = INIT_Y;
|
||
|
static uint32_t z = INIT_Z;
|
||
|
static uint32_t w = INIT_W;
|
||
|
uint32_t t = x ^ (x << 11);
|
||
|
x = y; y = z; z = w;
|
||
|
w ^= (w >> 19) ^ t ^ (t >> 8);
|
||
|
return w;
|
||
|
}
|
||
|
|
||
|
void test_cmp(uint32_t a, uint32_t b)
|
||
|
{
|
||
|
ezMiniSAT sat;
|
||
|
|
||
|
printf("A = %10u (%10d)\n", a, int32_t(a));
|
||
|
printf("B = %10u (%10d)\n", b, int32_t(b));
|
||
|
printf("\n");
|
||
|
|
||
|
std::vector<int> va = sat.vec_var("a", 32);
|
||
|
std::vector<int> vb = sat.vec_var("b", 32);
|
||
|
|
||
|
sat.vec_set_unsigned(va, a);
|
||
|
sat.vec_set_unsigned(vb, b);
|
||
|
|
||
|
#define MONITOR_VARS \
|
||
|
X(carry) X(overflow) X(sign) X(zero) \
|
||
|
X(lt_signed) X(le_signed) X(ge_signed) X(gt_signed) \
|
||
|
X(lt_unsigned) X(le_unsigned) X(ge_unsigned) X(gt_unsigned)
|
||
|
|
||
|
#define X(_n) int _n; bool _n ## _master;
|
||
|
MONITOR_VARS
|
||
|
#undef X
|
||
|
|
||
|
carry_master = ((uint64_t(a) - uint64_t(b)) >> 32) & 1;
|
||
|
overflow_master = (int32_t(a) - int32_t(b)) != (int64_t(int32_t(a)) - int64_t(int32_t(b)));
|
||
|
sign_master = ((a - b) >> 31) & 1;
|
||
|
zero_master = a == b;
|
||
|
|
||
|
sat.vec_cmp(va, vb, carry, overflow, sign, zero);
|
||
|
|
||
|
lt_signed_master = int32_t(a) < int32_t(b);
|
||
|
le_signed_master = int32_t(a) <= int32_t(b);
|
||
|
ge_signed_master = int32_t(a) >= int32_t(b);
|
||
|
gt_signed_master = int32_t(a) > int32_t(b);
|
||
|
|
||
|
lt_unsigned_master = a < b;
|
||
|
le_unsigned_master = a <= b;
|
||
|
ge_unsigned_master = a >= b;
|
||
|
gt_unsigned_master = a > b;
|
||
|
|
||
|
lt_signed = sat.vec_lt_signed(va, vb);
|
||
|
le_signed = sat.vec_le_signed(va, vb);
|
||
|
ge_signed = sat.vec_ge_signed(va, vb);
|
||
|
gt_signed = sat.vec_gt_signed(va, vb);
|
||
|
|
||
|
lt_unsigned = sat.vec_lt_unsigned(va, vb);
|
||
|
le_unsigned = sat.vec_le_unsigned(va, vb);
|
||
|
ge_unsigned = sat.vec_ge_unsigned(va, vb);
|
||
|
gt_unsigned = sat.vec_gt_unsigned(va, vb);
|
||
|
|
||
|
std::vector<int> modelExpressions;
|
||
|
std::vector<bool> modelValues, modelMaster;
|
||
|
std::vector<std::string> modelNames;
|
||
|
|
||
|
#define X(_n) modelExpressions.push_back(_n); modelNames.push_back(#_n); modelMaster.push_back(_n ## _master);
|
||
|
MONITOR_VARS
|
||
|
#undef X
|
||
|
|
||
|
std::vector<int> add_ab = sat.vec_add(va, vb);
|
||
|
std::vector<int> sub_ab = sat.vec_sub(va, vb);
|
||
|
std::vector<int> sub_ba = sat.vec_sub(vb, va);
|
||
|
|
||
|
sat.vec_append(modelExpressions, add_ab);
|
||
|
sat.vec_append(modelExpressions, sub_ab);
|
||
|
sat.vec_append(modelExpressions, sub_ba);
|
||
|
|
||
|
if (!sat.solve(modelExpressions, modelValues)) {
|
||
|
fprintf(stderr, "SAT solver failed to find a model!\n");
|
||
|
abort();
|
||
|
}
|
||
|
|
||
|
bool found_error = false;
|
||
|
|
||
|
for (size_t i = 0; i < modelMaster.size(); i++) {
|
||
|
if (modelMaster.at(i) != int(modelValues.at(i)))
|
||
|
found_error = true;
|
||
|
printf("%-20s %d%s\n", modelNames.at(i).c_str(), int(modelValues.at(i)),
|
||
|
modelMaster.at(i) != modelValues.at(i) ? " !!!" : "");
|
||
|
}
|
||
|
printf("\n");
|
||
|
|
||
|
uint32_t add_ab_value = sat.vec_model_get_unsigned(modelExpressions, modelValues, add_ab);
|
||
|
uint32_t sub_ab_value = sat.vec_model_get_unsigned(modelExpressions, modelValues, sub_ab);
|
||
|
uint32_t sub_ba_value = sat.vec_model_get_unsigned(modelExpressions, modelValues, sub_ba);
|
||
|
|
||
|
printf("%-20s %10u %10u%s\n", "result(a+b)", add_ab_value, a+b, add_ab_value != a+b ? " !!!" : "");
|
||
|
printf("%-20s %10u %10u%s\n", "result(a-b)", sub_ab_value, a-b, sub_ab_value != a-b ? " !!!" : "");
|
||
|
printf("%-20s %10u %10u%s\n", "result(b-a)", sub_ba_value, b-a, sub_ba_value != b-a ? " !!!" : "");
|
||
|
printf("\n");
|
||
|
|
||
|
if (found_error || add_ab_value != a+b || sub_ab_value != a-b || sub_ba_value != b-a)
|
||
|
abort();
|
||
|
}
|
||
|
|
||
|
int main()
|
||
|
{
|
||
|
printf("\n");
|
||
|
for (int i = 0; i < 1024; i++) {
|
||
|
printf("************** %d **************\n\n", i);
|
||
|
uint32_t a = xorshift128();
|
||
|
uint32_t b = xorshift128();
|
||
|
if (xorshift128() % 16 == 0)
|
||
|
a = b;
|
||
|
test_cmp(a, b);
|
||
|
}
|
||
|
return 0;
|
||
|
}
|
||
|
|