dnomd343
/
klotski
mirror of https://github.com/dnomd343/klotski.git
1
0
Fork 0
Code Issues Projects Releases Wiki Activity
Browse Source

perf: optimize range generation

legacy
Dnomd343 7 months ago
parent
6e49a2d57d
commit
f9a88410be
5 changed files with 225 additions and 44 deletions
Whitespace
Show all changes Ignore whitespace when comparing lines Ignore changes in amount of whitespace Ignore changes in whitespace at EOL
Split View
Diff Options
Show Stats Download Patch File Download Diff File
  1. 66
      src/core/benchmark/group.cc
  2. 2
      src/core/group/group.h
  3. 128
      src/core/group/internal/group.cc
  4. 33
      src/core/main.cc
  5. 40
      src/core_test/group_tmp/group_extend.cc

66
src/core/benchmark/group.cc
View File

@ -109,9 +109,73 @@ static void GroupExtend(benchmark::State &state) {
}
// static void FilterFromAllCases(benchmark::State &state) {
//
// klotski::cases::AllCases::instance().build();
//
// for (auto _ : state) {
// for (uint64_t head = 0; head < 16; ++head) {
//
// for (const auto range : AllCases::instance().fetch()[head]) {
// uint64_t common_code = head << 32 | range;
//
// volatile auto ret = klotski::cases::common_code_to_type_id(common_code);
//
// }
//
// }
// }
//
// }
static std::vector<std::tuple<int, int, int, int>> target_nums() {
std::vector<std::tuple<int, int, int, int>> results;
for (int n = 0; n <= 7; ++n) {
for (int n_2x1 = 0; n_2x1 <= n; ++n_2x1) {
for (int n_1x1 = 0; n_1x1 <= (14 - n * 2); ++n_1x1) {
results.emplace_back(16 - n*2 - n_1x1, n - n_2x1, n_2x1, n_1x1);
}
}
}
results.resize(203);
return results;
}
static void SpawnRanges(benchmark::State &state) {
auto nums = target_nums();
// std::cout << nums.size() << std::endl;
//
// for (auto [n1, n2, n3, n4] : nums) {
// if (n1 == 2 && n2 == 1 && n3 == 4 && n4 == 4) {
//
// std::cout << "ok" << std::endl;
//
// }
// }
for (auto _ : state) {
// klotski::cases::spawn_ranges(2, 1, 4, 4);
for (auto [n1, n2, n3, n4] : nums) {
klotski::cases::spawn_ranges(n1, n2, n3, n4);
}
}
}
// BENCHMARK(CommonCodeToTypeId)->Arg(8)->Arg(64)->Arg(256);
// BENCHMARK(RawCodeToTypeId)->Arg(8)->Arg(64)->Arg(256);
BENCHMARK(GroupExtend)->Unit(benchmark::kMillisecond);
// BENCHMARK(GroupExtend)->Unit(benchmark::kMillisecond);
// BENCHMARK(FilterFromAllCases)->Unit(benchmark::kMillisecond);
BENCHMARK(SpawnRanges)->Unit(benchmark::kMillisecond);
BENCHMARK_MAIN();

2
src/core/group/group.h
View File

@ -80,6 +80,8 @@ uint32_t raw_code_to_type_id(uint64_t raw_code);
std::vector<uint64_t> group_extend_from_seed(uint64_t raw_code);
std::vector<uint32_t> spawn_ranges(int n_00, int n_01, int n_10, int n_11);
class Group;
// TODO: add constexpr

128
src/core/group/internal/group.cc
View File

@ -2,10 +2,10 @@
#include <queue>
#include <absl/container/btree_map.h>
#include <absl/container/btree_set.h>
#include <absl/container/flat_hash_map.h>
#include <absl/container/node_hash_map.h>
#include <core/core.h>
static KLSK_INLINE uint32_t type_id(const int n, const int n_2x1, const int n_1x1) {
@ -31,78 +31,126 @@ std::vector<uint64_t> klotski::cases::group_extend_from_seed(uint64_t raw_code)
auto max_size = GroupUnion::create(raw_code_to_type_id(raw_code))->max_group_size();
// std::queue<uint64_t> cache({raw_code});
uint64_t offset = 0;
std::vector<uint64_t> results;
results.reserve(max_size);
results.emplace_back(raw_code);
// uint64_t offset = 0;
// std::vector<std::pair<uint64_t, uint64_t>> results;
// results.reserve(max_size);
// results.emplace_back(raw_code, 0);
absl::flat_hash_map<uint64_t, uint64_t> cases; // <code, mask>
// absl::node_hash_map<uint64_t, uint64_t> cases; // <code, mask>
// std::unordered_map<uint64_t, uint64_t> cases; // <code, mask>
// std::map<uint64_t, uint64_t> cases; // <code, mask>
// absl::btree_map<uint64_t, uint64_t> cases; // <code, mask>
cases.reserve(max_size);
cases.emplace(raw_code, 0); // without mask
// std::cout << max_size << std::endl;
auto core = klotski::core::Core(
[&results, &cases](auto code, auto mask) { // callback function
auto current = cases.find(code);
if (current != cases.end()) {
current->second |= mask; // update mask
// results[current->second].second |= mask; // update mask
return;
}
cases.emplace(code, mask);
// cases.emplace(code, results.size());
// cache.emplace(code);
results.emplace_back(code);
// cache.emplace(code, 0);
// results.emplace_back(code, 0);
}
);
while (offset != results.size()) {
auto tmp = results[offset];
core.next_cases(tmp, cases.find(tmp)->second);
// core.next_cases(tmp, 0);
// core.next_cases(tmp.first, tmp.second);
++offset;
}
return results;
// return {};
}
template<int N>
static std::vector<uint32_t> demo(int n_10, int n_11) {
// while (!cache.empty()) { // until BFS without elements
// core.next_cases(cache.front(), cases.find(cache.front())->second);
// cache.pop(); // case dequeue
// }
constexpr auto num = 16 - N;
constexpr auto offset = (16 - num) << 1; // offset of low bits
int n_00 = 16 - N * 2 - n_11;
int n_01 = N - n_10;
// std::vector<uint64_t> result;
// result.reserve(cases.size());
// for (auto &&tmp : cases) { // export group cases
// result.emplace_back(tmp.first);
std::array<int, num> series {};
// for (int k = 0; k < n_00; ++k) {
// series[k] = 0b00;
// }
// for (int k = n_00; k < n_00 + n_01; ++k) {
// series[k] = 0b01;
// }
// return result;
// std::vector<uint64_t> result;
// result.reserve(results.size());
// for (auto [code, _] : results) {
// result.emplace_back(code);
auto kk = std::fill_n(series.begin() + n_00, n_01, 0b01);
auto pp = std::fill_n(kk, n_10, 0b10);
std::fill_n(pp, n_11, 0b11);
// for (auto x : series) {
// std::cout << x << " ";
// }
// return result;
// std::cout << std::endl;
std::vector<uint32_t> ranges;
do {
uint32_t range = 0;
for (const auto x : series) // store every 2-bit
(range <<= 2) |= x;
ranges.emplace_back(range << offset);
} while (std::ranges::next_permutation(series).found);
return ranges;
}
std::vector<uint32_t> klotski::cases::spawn_ranges(int n_00, int n_01, int n_10, int n_11) {
auto n = n_01 + n_10;
switch (n) {
case 0: return demo<0>(n_10, n_11);
case 1: return demo<1>(n_10, n_11);
case 2: return demo<2>(n_10, n_11);
case 3: return demo<3>(n_10, n_11);
case 4: return demo<4>(n_10, n_11);
case 5: return demo<5>(n_10, n_11);
case 6: return demo<6>(n_10, n_11);
case 7: return demo<7>(n_10, n_11);
default: return {};
}
// return demo<5>(n_10, n_11);
// auto num = n_00 + n_01 + n_10 + n_11;
// auto offset = (16 - num) << 1; // offset of low bits
// std::vector<int> series;
// series.reserve(num);
// series.insert(series.end(), n_00, 0b00);
// series.insert(series.end(), n_01, 0b01);
// series.insert(series.end(), n_10, 0b10);
// series.insert(series.end(), n_11, 0b11);
// std::array<int, 11> series {
// 0b00, 0b00,
// 0b01,
// 0b10, 0b10, 0b10, 0b10,
// 0b11, 0b11, 0b11, 0b11,
// };
// std::vector<uint32_t> ranges;
// do { // full permutation traversal
// uint32_t range = 0;
// for (const auto x : series) // store every 2-bit
// (range <<= 2) |= x;
// ranges.emplace_back(range << offset);
// } while (std::next_permutation(series.begin(), series.end()));
// do {
// uint32_t range = 0;
// for (const auto x : series) // store every 2-bit
// (range <<= 2) |= x;
// ranges.emplace_back(range << offset);
// } while (std::ranges::next_permutation(series).found);
// return ranges;
}

33
src/core/main.cc
View File

@ -1,3 +1,4 @@
#include <algorithm>
#include <thread>
#include <iostream>
#include <format>
@ -25,10 +26,36 @@ using klotski::codec::SHORT_CODE_LIMIT;
int main() {
const auto start = clock();
auto raw_code = RawCode::from_common_code(0x1A9BF0C00)->unwrap();
auto ret = klotski::cases::group_extend_from_seed(raw_code);
klotski::cases::spawn_ranges(2, 1, 4, 4);
std::cout << ret.size() << std::endl;
// std::vector<int> series {1, 2, 3, 4};
// do { // full permutation traversal
//
// for (auto s : series) {
// std::cout << s << " ";
// }
// std::cout << std::endl;
//
// } while (std::next_permutation(series.begin(), series.end()));
// std::array a{'a', 'b', 'c'};
// do {
// for (auto x : a) {
// std::cout << x;
// }
// std::cout << std::endl;
// }
// while (std::ranges::next_permutation(a).found);
// auto raw_code = RawCode::from_common_code(0x1A9BF0C00)->unwrap();
// auto ret = klotski::cases::group_extend_from_seed(raw_code);
//
// std::cout << ret.size() << std::endl;
// 1 A9BF0C00 -> 10 10 10 01 10 11 11 11 00 00 11 000000
// auto ret = klotski::cases::spawn_ranges(2, 1, 4, 4);
// std::cout << ret.size() << std::endl;
// auto kk = GroupUnion::create(123).value();
// std::cout << kk.size() << std::endl;

40
src/core_test/group_tmp/group_extend.cc
View File

@ -20,3 +20,43 @@ TEST(Group, group_extend) {
auto hash_ret = hash::xxh3(codes.data(), codes.size() * sizeof(uint64_t));
EXPECT_EQ(hash_ret, 0x91BD28A749312A6D);
}
static std::vector<std::tuple<int, int, int, int>> target_nums() {
std::vector<std::tuple<int, int, int, int>> results;
for (int n = 0; n <= 7; ++n) {
for (int n_2x1 = 0; n_2x1 <= n; ++n_2x1) {
for (int n_1x1 = 0; n_1x1 <= (14 - n * 2); ++n_1x1) {
results.emplace_back(16 - n*2 - n_1x1, n - n_2x1, n_2x1, n_1x1);
}
}
}
// results.resize(203);
return results;
}
TEST(Group, ranges) {
// auto ret = klotski::cases::spawn_ranges(2, 1, 4, 4);
//
// EXPECT_EQ(ret.size(), 34650);
//
// auto hash_ret = hash::xxh3(ret.data(), ret.size() * 4);
// EXPECT_EQ(hash_ret, 0xF6F87606E4205EAF);
std::vector<uint32_t> ranges;
for (auto [n1, n2, n3, n4] : target_nums()) {
auto kk = klotski::cases::spawn_ranges(n1, n2, n3, n4);
ranges.insert(ranges.end(), kk.begin(), kk.end());
}
EXPECT_EQ(ranges.size(), 7311921);
auto hash_ret = hash::xxh3(ranges.data(), ranges.size() * 4);
EXPECT_EQ(hash_ret, 0xA1E247B01D5A9545);
}

Write Preview
Loading…
Cancel
Save