Skip to content

Use pod_vector for mmap zero-copy support in serialized data structures #12

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Open
adamant-pwn wants to merge 1 commit into
base: main
Choose a base branch
from
+33 −29
Open

Use pod_vector for mmap zero-copy support in serialized data structures #12

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
2 changes: 1 addition & 1 deletion external/essentials
Open in desktop
Submodule essentials updated 1 files
+182 −55 include/essentials.hpp
6 changes: 3 additions & 3 deletions include/bit_vector.hpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -41,7 +41,7 @@ struct bit_vector //

void build(bit_vector& bv) {
bv.m_num_bits = m_num_bits;
bv.m_data.swap(m_data);
bv.m_data = std::move(m_data);
builder().swap(*this);
}

Expand Down Expand Up @@ -322,7 +322,7 @@ struct bit_vector //
iterator begin() const { return get_iterator_at(0); }

uint64_t num_bits() const { return m_num_bits; }
std::vector<uint64_t> const& data() const { return m_data; }
essentials::owning_span<uint64_t> const& data() const { return m_data; }

uint64_t num_bytes() const { return sizeof(m_num_bits) + essentials::vec_bytes(m_data); }

Expand All @@ -343,7 +343,7 @@ struct bit_vector //

protected:
uint64_t m_num_bits;
std::vector<uint64_t> m_data;
essentials::owning_span<uint64_t> m_data;

template <typename Visitor, typename T>
static void visit_impl(Visitor& visitor, T&& t) {
Expand Down
8 changes: 4 additions & 4 deletions include/cache_line_elias_fano.hpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -52,10 +52,9 @@ struct cache_line_elias_fano {

const uint64_t num_blocks = (n + 44 - 1) / 44;
const uint64_t num_bytes = num_blocks * 64;
m_bits.resize(num_bytes);
std::fill(m_bits.begin(), m_bits.end(), 0);
std::vector<uint8_t> bits(num_bytes);

uint8_t* high = m_bits.data();
uint8_t* high = bits.data();
uint8_t* low = high + 4 // for lower bound high part of block
+ 16; // for high bits

Expand Down Expand Up @@ -100,6 +99,7 @@ struct cache_line_elias_fano {
}

m_back = last;
m_bits = essentials::owning_span<uint8_t>(std::move(bits));
}

uint64_t access(uint64_t i) const {
Expand Down Expand Up @@ -149,7 +149,7 @@ struct cache_line_elias_fano {
private:
uint64_t m_back;
uint64_t m_size;
std::vector<uint8_t> m_bits;
essentials::owning_span<uint8_t> m_bits;

template <typename Visitor, typename T>
static void visit_impl(Visitor& visitor, T&& t) {
Expand Down
6 changes: 3 additions & 3 deletions include/compact_vector.hpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -195,7 +195,7 @@ struct compact_vector //
cv.m_size = m_size;
cv.m_width = m_width;
cv.m_mask = m_mask;
cv.m_data.swap(m_data);
cv.m_data = std::move(m_data);
builder().swap(*this);
}

Expand Down Expand Up @@ -262,7 +262,7 @@ struct compact_vector //
uint64_t back() const { return operator[](size() - 1); }
uint64_t size() const { return m_size; }
uint64_t width() const { return m_width; }
std::vector<uint64_t> const& data() const { return m_data; }
essentials::owning_span<uint64_t> const& data() const { return m_data; }

typedef enumerator<compact_vector> iterator;
iterator get_iterator_at(uint64_t pos) const { return iterator(this, pos); }
Expand Down Expand Up @@ -293,7 +293,7 @@ struct compact_vector //
uint64_t m_size;
uint64_t m_width;
uint64_t m_mask;
std::vector<uint64_t> m_data;
essentials::owning_span<uint64_t> m_data;

template <typename Visitor, typename T>
static void visit_impl(Visitor& visitor, T&& t) {
Expand Down
22 changes: 12 additions & 10 deletions include/darray.hpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -81,7 +81,7 @@ struct darray {
darray() : m_positions(0) {}

void build(bit_vector const& B) {
std::vector<uint64_t> const& data = B.data();
auto const& data = B.data();
std::vector<uint64_t> cur_block_positions;
std::vector<int64_t> block_inventory;
std::vector<uint16_t> subblock_inventory;
Expand Down Expand Up @@ -113,9 +113,9 @@ struct darray {
flush_cur_block(cur_block_positions, block_inventory, subblock_inventory,
overflow_positions);
}
m_block_inventory.swap(block_inventory);
m_subblock_inventory.swap(subblock_inventory);
m_overflow_positions.swap(overflow_positions);
m_block_inventory = std::move(block_inventory);
m_subblock_inventory = std::move(subblock_inventory);
m_overflow_positions = std::move(overflow_positions);

// std::cout << "I: ";
// for (auto x : m_block_inventory) { std::cout << x << ' '; }
Expand Down Expand Up @@ -174,7 +174,7 @@ struct darray {
uint64_t reminder = i & (subblock_size - 1);
if (!reminder) return start_pos;

std::vector<uint64_t> const& data = B.data();
auto const& data = B.data();
uint64_t word_idx = start_pos >> 6;
uint64_t word_shift = start_pos & 63;
uint64_t word = WordGetter()(data, word_idx) & (uint64_t(-1) << word_shift);
Expand Down Expand Up @@ -214,9 +214,9 @@ struct darray {

protected:
uint64_t m_positions;
std::vector<int64_t> m_block_inventory;
std::vector<uint16_t> m_subblock_inventory;
std::vector<uint64_t> m_overflow_positions;
essentials::owning_span<int64_t> m_block_inventory;
essentials::owning_span<uint16_t> m_subblock_inventory;
essentials::owning_span<uint64_t> m_overflow_positions;

template <typename Visitor, typename T>
static void visit_impl(Visitor& visitor, T&& t) {
Expand Down Expand Up @@ -262,11 +262,13 @@ struct darray {
namespace util {

struct identity_getter {
uint64_t operator()(std::vector<uint64_t> const& data, uint64_t i) const { return data[i]; }
template <typename Vec>
uint64_t operator()(Vec const& data, uint64_t i) const { return data[i]; }
};

struct negating_getter {
uint64_t operator()(std::vector<uint64_t> const& data, uint64_t i) const { return ~data[i]; }
template <typename Vec>
uint64_t operator()(Vec const& data, uint64_t i) const { return ~data[i]; }
};

} // namespace util
Expand Down
10 changes: 6 additions & 4 deletions include/endpoints_sequence.hpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -50,7 +50,8 @@ struct endpoints_sequence {

const uint64_t num_high_bits = n + (universe >> 8) + 1;
bit_vector::builder bvb_high_bits(num_high_bits);
m_low_bits.reserve(n);
std::vector<uint8_t> low_bits;
low_bits.reserve(n);

compact_vector::builder cvb_hints_0((universe + 256 - 1) / 256, // ceil(U/2^8)
util::ceil_log2_uint64(num_high_bits));
Expand All @@ -59,7 +60,7 @@ struct endpoints_sequence {
uint64_t prev_pos = 0;
for (uint64_t i = 0; i != n; ++i, ++begin) {
auto v = *begin;
m_low_bits.push_back(v & 255);
low_bits.push_back(v & 255);
uint64_t high_part = v >> 8;
uint64_t pos = high_part + i;
bvb_high_bits.set(pos, 1);
Expand All @@ -72,6 +73,7 @@ struct endpoints_sequence {
bvb_high_bits.build(m_high_bits);
m_high_bits_d1.build(m_high_bits);
cvb_hints_0.build(m_hints_0);
m_low_bits = essentials::owning_span<uint8_t>(std::move(low_bits));
}

struct iterator {
Expand Down Expand Up @@ -141,7 +143,7 @@ struct endpoints_sequence {
uint64_t m_pos;
uint64_t m_val;
bit_vector::iterator m_high_bits_it;
std::vector<uint8_t>::const_iterator m_low_bits_it;
const uint8_t* m_low_bits_it;

void read_next_value() {
assert(m_pos < m_ptr->size());
Expand Down Expand Up @@ -226,7 +228,7 @@ struct endpoints_sequence {
bit_vector m_high_bits;
DArray1 m_high_bits_d1;
compact_vector m_hints_0;
std::vector<uint8_t> m_low_bits;
essentials::owning_span<uint8_t> m_low_bits;

template <typename Visitor, typename T>
static void visit_impl(Visitor& visitor, T&& t) {
Expand Down
8 changes: 4 additions & 4 deletions include/rank9.hpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -46,7 +46,7 @@ struct rank9 {
rank9() {}

void build(bit_vector const& B) {
std::vector<uint64_t> const& data = B.data();
auto const& data = B.data();
std::vector<uint64_t> block_rank_pairs;
uint64_t next_rank = 0;
uint64_t cur_subrank = 0;
Expand Down Expand Up @@ -81,7 +81,7 @@ struct rank9 {
block_rank_pairs.push_back(0);
}

m_block_rank_pairs.swap(block_rank_pairs);
m_block_rank_pairs = std::move(block_rank_pairs);
}

inline uint64_t num_ones() const { return *(m_block_rank_pairs.end() - 2); }
Expand All @@ -96,7 +96,7 @@ struct rank9 {
uint64_t r = sub_block_rank(sub_block);
uint64_t sub_left = i % 64;
if (sub_left) {
std::vector<uint64_t> const& data = B.data();
auto const& data = B.data();
r += util::popcount(data[sub_block] << (64 - sub_left));
}
return r;
Expand Down Expand Up @@ -146,7 +146,7 @@ struct rank9 {
}

static const uint64_t block_size = 8; // in 64bit words
std::vector<uint64_t> m_block_rank_pairs;
essentials::owning_span<uint64_t> m_block_rank_pairs;
};

} // namespace bits