initial runs
This commit is contained in:
parent
9e9260c84d
commit
918c1cba78
18 changed files with 379 additions and 356 deletions
BIN
.DS_Store
vendored
BIN
.DS_Store
vendored
Binary file not shown.
5
.clangd
5
.clangd
|
|
@ -1,2 +1,5 @@
|
|||
CompileFlags:
|
||||
Add: [-std=c++17]
|
||||
Add:
|
||||
- "-std=c++20"
|
||||
- "-I./include"
|
||||
CompilationDatabase: build
|
||||
|
|
|
|||
|
|
@ -1,69 +0,0 @@
|
|||
#pragma once
|
||||
|
||||
#include "record.hpp"
|
||||
#include <stdexcept>
|
||||
#include <vector>
|
||||
|
||||
class Buffer;
|
||||
|
||||
class SubBuffer {
|
||||
private:
|
||||
Record* data_;
|
||||
size_t size_;
|
||||
|
||||
friend class Buffer;
|
||||
|
||||
SubBuffer(Record* data, size_t size)
|
||||
: data_(data)
|
||||
, size_(size)
|
||||
{
|
||||
}
|
||||
|
||||
public:
|
||||
// so that i can create a vecotr of set size, without the .reserve(), and other stuff
|
||||
SubBuffer()
|
||||
{
|
||||
}
|
||||
// for std::sort support
|
||||
using iterator = Record*;
|
||||
iterator begin() { return data_; }
|
||||
iterator end() { return data_ + size_; }
|
||||
|
||||
Record& operator[](size_t index)
|
||||
{
|
||||
return data_[index];
|
||||
}
|
||||
|
||||
size_t size() const { return size_; }
|
||||
};
|
||||
|
||||
class Buffer {
|
||||
private:
|
||||
std::vector<Record> storage_;
|
||||
|
||||
public:
|
||||
explicit Buffer(size_t total_size)
|
||||
: storage_(total_size)
|
||||
{
|
||||
}
|
||||
|
||||
std::vector<SubBuffer> divide(size_t n)
|
||||
{
|
||||
if (n == 0) {
|
||||
throw std::invalid_argument("Cannot divide buffer into 0 pieces");
|
||||
}
|
||||
|
||||
size_t total_size = storage_.size();
|
||||
if (total_size % n != 0) {
|
||||
throw std::invalid_argument("Size not divisible by n");
|
||||
}
|
||||
|
||||
std::vector<SubBuffer> result(n);
|
||||
size_t piece_size = total_size / n;
|
||||
for (size_t i = 0; i < n; ++i) {
|
||||
result[i] = SubBuffer(storage_.data() + i * piece_size, piece_size);
|
||||
}
|
||||
|
||||
return result;
|
||||
}
|
||||
};
|
||||
|
|
@ -1,70 +1,17 @@
|
|||
#pragma once
|
||||
#include <getopt.h>
|
||||
#include <iostream>
|
||||
#include <stdexcept>
|
||||
#include <string>
|
||||
|
||||
class Configuration {
|
||||
public:
|
||||
std::string input_file = "1.in";
|
||||
std::string output_file = "1.in";
|
||||
std::string directory = "tmp";
|
||||
std::string output_file = "1.out";
|
||||
std::string tmp_dir = "build/run";
|
||||
bool generate_data = false;
|
||||
bool evaluate_file = false;
|
||||
int N = 100000; // number of records in a file
|
||||
int n = 101; // number of buffers
|
||||
int b = 10; // blocking factor
|
||||
|
||||
static Configuration parse_args(int argc, char** argv)
|
||||
{
|
||||
if (argc <= 1) {
|
||||
std::cerr << "Please specify input arguments\n";
|
||||
std::cerr << "Usage: " << argv[0] << " [-g] [-e] [-N count] [-i input] [-o output] [-d directory] [-n buffers] [-b block_size]\n";
|
||||
throw std::invalid_argument("No input arguments");
|
||||
}
|
||||
Configuration config;
|
||||
bool input_set = false;
|
||||
int opt;
|
||||
|
||||
while ((opt = getopt(argc, argv, "i:o:d:geN:n:b:")) != -1) {
|
||||
switch (opt) {
|
||||
case 'i':
|
||||
config.input_file = optarg;
|
||||
input_set = true;
|
||||
break;
|
||||
case 'o':
|
||||
config.output_file = optarg;
|
||||
break;
|
||||
case 'd':
|
||||
config.directory = optarg;
|
||||
break;
|
||||
case 'g':
|
||||
config.generate_data = true;
|
||||
break;
|
||||
case 'e':
|
||||
config.evaluate_file = true;
|
||||
break;
|
||||
case 'N':
|
||||
config.N = std::stoi(optarg);
|
||||
break;
|
||||
case 'n':
|
||||
config.n = std::stoi(optarg);
|
||||
break;
|
||||
case 'b':
|
||||
config.b = std::stoi(optarg);
|
||||
break;
|
||||
case '?':
|
||||
throw std::invalid_argument("Invalid command line argument");
|
||||
}
|
||||
}
|
||||
|
||||
if (config.generate_data && input_set) {
|
||||
throw std::invalid_argument("Cannot specify both -g and -i options");
|
||||
}
|
||||
if (config.generate_data && config.evaluate_file) {
|
||||
throw std::invalid_argument("Cannot specify both -g and -e options");
|
||||
}
|
||||
|
||||
return config;
|
||||
}
|
||||
Configuration(int argc, char** argv);
|
||||
};
|
||||
|
|
|
|||
22
include/file_reader.hpp
Normal file
22
include/file_reader.hpp
Normal file
|
|
@ -0,0 +1,22 @@
|
|||
#pragma once
|
||||
#include "record.hpp"
|
||||
#include <fstream>
|
||||
#include <optional>
|
||||
#include <span>
|
||||
|
||||
class FileReader {
|
||||
private:
|
||||
std::ifstream stream_;
|
||||
std::vector<Record> buffer_;
|
||||
size_t current_pos_;
|
||||
size_t valid_items_;
|
||||
size_t total_reads_;
|
||||
|
||||
void refill_buffer();
|
||||
|
||||
public:
|
||||
FileReader(const std::string& filename, size_t buffer_size = 8);
|
||||
std::optional<Record> read();
|
||||
size_t read_chunk(std::span<Record>& chunk);
|
||||
size_t total_reads() const;
|
||||
};
|
||||
23
include/file_writer.hpp
Normal file
23
include/file_writer.hpp
Normal file
|
|
@ -0,0 +1,23 @@
|
|||
#pragma once
|
||||
#include "record.hpp"
|
||||
#include <fstream>
|
||||
#include <span>
|
||||
|
||||
class FileWriter {
|
||||
private:
|
||||
std::ofstream stream_;
|
||||
std::vector<Record> buffer_;
|
||||
size_t current_pos_;
|
||||
size_t total_writes_;
|
||||
|
||||
void write_buffer();
|
||||
|
||||
public:
|
||||
FileWriter(const std::string& filename, size_t buffer_size = 8);
|
||||
~FileWriter();
|
||||
|
||||
void write(const Record& record);
|
||||
void write_chunk(std::span<const Record>& chunk);
|
||||
|
||||
size_t total_writes() const;
|
||||
};
|
||||
29
include/merge_sorter.hpp
Normal file
29
include/merge_sorter.hpp
Normal file
|
|
@ -0,0 +1,29 @@
|
|||
#pragma once
|
||||
#include "file_reader.hpp"
|
||||
#include "file_writer.hpp"
|
||||
#include "record.hpp"
|
||||
#include <vector>
|
||||
|
||||
class MergeSorter {
|
||||
private:
|
||||
std::vector<Record> buffer_;
|
||||
const std::string& tmp_dir_;
|
||||
size_t buffer_rows = 0;
|
||||
size_t disk_reads_ = 0;
|
||||
size_t disk_writes_ = 0;
|
||||
size_t phases_ = 0;
|
||||
|
||||
size_t create_initial_runs(const std::string& input_file);
|
||||
void merge_runs(const std::vector<std::string>& input_files,
|
||||
const std::string& output_file);
|
||||
void perform_k_way_merge(std::vector<std::unique_ptr<FileReader>>& readers,
|
||||
FileWriter& writer);
|
||||
|
||||
public:
|
||||
MergeSorter(size_t buffer_rows, size_t buffer_cols, const std::string& tmp_dir);
|
||||
bool sort_file(const std::string& input_file, const std::string& output_file);
|
||||
|
||||
size_t disk_reads() const;
|
||||
size_t disk_writes() const;
|
||||
size_t phases() const;
|
||||
};
|
||||
|
|
@ -1,79 +0,0 @@
|
|||
// #todo maybe overload >> operators insted of read()
|
||||
#pragma once
|
||||
#include "record.hpp"
|
||||
#include <istream>
|
||||
#include <optional>
|
||||
#include <vector>
|
||||
|
||||
class Reader {
|
||||
private:
|
||||
std::vector<Record> read_buffer;
|
||||
std::istream& input_stream;
|
||||
size_t buffer_size;
|
||||
size_t total_reads_count;
|
||||
int last_element_idx; // last valid element in buffer
|
||||
int current_idx; // current reading position
|
||||
|
||||
// returns true if at least one record was returned
|
||||
bool refill_buffer()
|
||||
{
|
||||
last_element_idx = -1;
|
||||
current_idx = 0;
|
||||
|
||||
for (size_t i = 0; i < buffer_size; i++) {
|
||||
if (!(input_stream >> read_buffer[i])) {
|
||||
break;
|
||||
}
|
||||
last_element_idx = i;
|
||||
}
|
||||
|
||||
return last_element_idx >= 0;
|
||||
}
|
||||
|
||||
public:
|
||||
explicit Reader(std::istream& input_stream, size_t buffer_size = 8)
|
||||
: read_buffer(buffer_size)
|
||||
, input_stream(input_stream)
|
||||
, buffer_size(buffer_size)
|
||||
, total_reads_count(0)
|
||||
, last_element_idx(-1)
|
||||
, current_idx(0)
|
||||
{
|
||||
}
|
||||
|
||||
std::optional<Record> read()
|
||||
{
|
||||
if (current_idx > last_element_idx) {
|
||||
bool read_new_items = refill_buffer();
|
||||
if (!read_new_items) {
|
||||
return std::nullopt;
|
||||
}
|
||||
}
|
||||
|
||||
Record record = read_buffer[current_idx];
|
||||
current_idx++;
|
||||
total_reads_count++;
|
||||
|
||||
return record;
|
||||
}
|
||||
|
||||
size_t total_reads() const
|
||||
{
|
||||
return total_reads_count;
|
||||
}
|
||||
|
||||
// #todo remove if not used in the end
|
||||
// bool has_more() const
|
||||
// {
|
||||
// return current_idx <= last_element_idx || !input_stream.eof();
|
||||
// }
|
||||
|
||||
// void reset()
|
||||
// {
|
||||
// input_stream.clear();
|
||||
// input_stream.seekg(0);
|
||||
// last_element_idx = -1;
|
||||
// current_idx = 0;
|
||||
// total_reads_count = 0;
|
||||
// }
|
||||
};
|
||||
|
|
@ -1,53 +1,17 @@
|
|||
#pragma once
|
||||
#include <array>
|
||||
#include <iostream>
|
||||
#include <random>
|
||||
|
||||
class Record {
|
||||
public:
|
||||
std::array<int, 5> a;
|
||||
int x;
|
||||
|
||||
int evaluate() const
|
||||
{
|
||||
int out = 0;
|
||||
int x_n = 1;
|
||||
for (int i = 0; i < 5; i++) {
|
||||
out += a[i] * x_n;
|
||||
x_n *= x;
|
||||
}
|
||||
return out;
|
||||
}
|
||||
int evaluate() const;
|
||||
bool operator<(const Record& other) const;
|
||||
|
||||
bool operator<(const Record& other) const
|
||||
{
|
||||
return evaluate() < other.evaluate();
|
||||
}
|
||||
friend std::ostream& operator<<(std::ostream& os, const Record& r);
|
||||
friend std::istream& operator>>(std::istream& is, Record& r);
|
||||
|
||||
friend std::ostream& operator<<(std::ostream& os, const Record& r)
|
||||
{
|
||||
os << r.a[0] << ' ' << r.a[1] << ' ' << r.a[2] << ' '
|
||||
<< r.a[3] << ' ' << r.a[4] << ' ' << r.x;
|
||||
return os;
|
||||
}
|
||||
|
||||
friend std::istream& operator>>(std::istream& is, Record& r)
|
||||
{
|
||||
is >> r.a[0] >> r.a[1] >> r.a[2] >> r.a[3] >> r.a[4] >> r.x;
|
||||
return is;
|
||||
}
|
||||
|
||||
static Record random()
|
||||
{
|
||||
static std::random_device rd;
|
||||
static std::mt19937 gen(rd());
|
||||
static std::uniform_int_distribution<int> dist;
|
||||
|
||||
Record r;
|
||||
for (int& val : r.a) {
|
||||
val = dist(gen);
|
||||
}
|
||||
r.x = dist(gen);
|
||||
return r;
|
||||
}
|
||||
};
|
||||
static Record random();
|
||||
};
|
||||
|
|
|
|||
|
|
@ -1,63 +0,0 @@
|
|||
// #todo maybe overload << operators insted of write()
|
||||
#pragma once
|
||||
#include "record.hpp"
|
||||
#include <ostream>
|
||||
#include <vector>
|
||||
|
||||
class Writer {
|
||||
private:
|
||||
std::vector<Record> write_buffer;
|
||||
std::ostream& output_stream;
|
||||
size_t buffer_size;
|
||||
size_t total_writes_count;
|
||||
size_t current_index;
|
||||
|
||||
// returns true on success
|
||||
// doesn't really flush for performace reasons
|
||||
// could be changed to really flush to monitor the impact of amount of writes
|
||||
bool flush_buffer()
|
||||
{
|
||||
for (size_t i = 0; i < current_index; ++i) {
|
||||
if (!(output_stream << write_buffer[i] << " ")) {
|
||||
return false;
|
||||
}
|
||||
}
|
||||
|
||||
current_index = 0;
|
||||
|
||||
return true;
|
||||
}
|
||||
|
||||
public:
|
||||
explicit Writer(std::ostream& output_stream, size_t buffer_size = 8)
|
||||
: write_buffer(buffer_size)
|
||||
, output_stream(output_stream)
|
||||
, buffer_size(buffer_size)
|
||||
, total_writes_count(0)
|
||||
, current_index(0)
|
||||
{
|
||||
}
|
||||
|
||||
~Writer()
|
||||
{
|
||||
flush_buffer();
|
||||
}
|
||||
|
||||
bool write(const Record& record)
|
||||
{
|
||||
write_buffer[current_index] = record;
|
||||
current_index++;
|
||||
total_writes_count++;
|
||||
|
||||
if (current_index >= buffer_size) {
|
||||
return flush_buffer();
|
||||
}
|
||||
|
||||
return true;
|
||||
}
|
||||
|
||||
size_t total_writes() const
|
||||
{
|
||||
return total_writes_count;
|
||||
}
|
||||
};
|
||||
15
makefile
15
makefile
|
|
@ -1,11 +1,18 @@
|
|||
CXX = clang++
|
||||
CFLAGS = -g -lc++ -std=c++17
|
||||
CXXFLAGS = -g -std=c++20 -Iinclude -Wall -Wextra -Werror
|
||||
LDFLAGS =
|
||||
|
||||
SRC := $(wildcard src/*.cpp)
|
||||
OBJ := $(SRC:src/%.cpp=build/%.o)
|
||||
TARGET = build/main
|
||||
|
||||
$(TARGET): $(SRC) | build
|
||||
$(CXX) $(CXXFLAGS) $(SRC) -o $(TARGET)
|
||||
all: $(TARGET)
|
||||
|
||||
$(TARGET): $(OBJ)
|
||||
$(CXX) $(CXXFLAGS) $^ -o $@ $(LDFLAGS)
|
||||
|
||||
build/%.o: src/%.cpp | build
|
||||
$(CXX) $(CXXFLAGS) -c $< -o $@
|
||||
|
||||
build:
|
||||
mkdir -p build build/run
|
||||
|
|
@ -25,4 +32,4 @@ run: $(TARGET) build/run/1.in
|
|||
run_alt: $(TARGET) build/run/1.in
|
||||
$(TARGET) -i build/run/1.in -d build/run/tmp -n 11 -b 100
|
||||
|
||||
.PHONY: clean run generate run_alt
|
||||
.PHONY: clean run generate run_alt all
|
||||
|
|
|
|||
55
src/config.cpp
Normal file
55
src/config.cpp
Normal file
|
|
@ -0,0 +1,55 @@
|
|||
#include "../include/config.hpp"
|
||||
#include <iostream>
|
||||
#include <stdexcept>
|
||||
|
||||
Configuration::Configuration(int argc, char** argv)
|
||||
{
|
||||
if (argc <= 1) {
|
||||
std::cerr << "Please specify input arguments\n";
|
||||
std::cerr << "Usage: " << argv[0] << " [-g] [-e] [-N count] [-i input] [-o output] [-d dir] [-n buffers] [-b block_size]\n";
|
||||
throw std::invalid_argument("No input arguments");
|
||||
}
|
||||
bool input_set = false;
|
||||
int opt;
|
||||
|
||||
while ((opt = getopt(argc, argv, "i:d:o:geN:n:b:")) != -1) {
|
||||
switch (opt) {
|
||||
case 'i':
|
||||
input_file = optarg;
|
||||
input_set = true;
|
||||
break;
|
||||
case 'o':
|
||||
output_file = optarg;
|
||||
break;
|
||||
case 'd':
|
||||
tmp_dir = optarg;
|
||||
break;
|
||||
case 'g':
|
||||
generate_data = true;
|
||||
break;
|
||||
case 'e':
|
||||
evaluate_file = true;
|
||||
break;
|
||||
case 'N':
|
||||
N = std::stoi(optarg);
|
||||
break;
|
||||
case 'n':
|
||||
n = std::stoi(optarg);
|
||||
break;
|
||||
case 'b':
|
||||
b = std::stoi(optarg);
|
||||
break;
|
||||
case '?':
|
||||
throw std::invalid_argument("Invalid command line argument");
|
||||
default:
|
||||
std::cerr << "Please specify input arguments\n";
|
||||
}
|
||||
}
|
||||
|
||||
if (generate_data && input_set) {
|
||||
throw std::invalid_argument("Cannot specify both -g and -i options");
|
||||
}
|
||||
if (generate_data && evaluate_file) {
|
||||
throw std::invalid_argument("Cannot specify both -g and -e options");
|
||||
}
|
||||
}
|
||||
62
src/file_reader.cpp
Normal file
62
src/file_reader.cpp
Normal file
|
|
@ -0,0 +1,62 @@
|
|||
#include "../include/file_reader.hpp"
|
||||
|
||||
void FileReader::refill_buffer()
|
||||
{
|
||||
current_pos_ = 0;
|
||||
valid_items_ = 0;
|
||||
|
||||
for (size_t i = 0; i < buffer_.size(); ++i) {
|
||||
Record record;
|
||||
if (!(stream_ >> record)) {
|
||||
break;
|
||||
}
|
||||
buffer_[i] = record;
|
||||
valid_items_ = i + 1;
|
||||
}
|
||||
}
|
||||
|
||||
FileReader::FileReader(const std::string& filename, size_t buffer_size)
|
||||
: stream_(filename)
|
||||
, buffer_(buffer_size)
|
||||
, current_pos_(0)
|
||||
, valid_items_(0)
|
||||
, total_reads_(0)
|
||||
{
|
||||
refill_buffer();
|
||||
}
|
||||
|
||||
std::optional<Record> FileReader::read()
|
||||
{
|
||||
if (current_pos_ >= valid_items_) {
|
||||
refill_buffer();
|
||||
}
|
||||
|
||||
if (valid_items_ == 0) {
|
||||
return std::nullopt;
|
||||
}
|
||||
|
||||
Record record = buffer_[current_pos_];
|
||||
current_pos_++;
|
||||
total_reads_++;
|
||||
|
||||
return record;
|
||||
}
|
||||
|
||||
size_t FileReader::read_chunk(std::span<Record>& chunk)
|
||||
{
|
||||
size_t count = 0;
|
||||
for (size_t i = 0; i < chunk.size(); i++) {
|
||||
auto record = read();
|
||||
if (!record.has_value()) {
|
||||
break;
|
||||
}
|
||||
chunk[count] = record.value();
|
||||
count++;
|
||||
}
|
||||
return count;
|
||||
}
|
||||
|
||||
size_t FileReader::total_reads() const
|
||||
{
|
||||
return total_reads_;
|
||||
}
|
||||
47
src/file_writer.cpp
Normal file
47
src/file_writer.cpp
Normal file
|
|
@ -0,0 +1,47 @@
|
|||
#include "../include/file_writer.hpp"
|
||||
|
||||
void FileWriter::write_buffer()
|
||||
{
|
||||
for (size_t i = 0; i < current_pos_; i++) {
|
||||
stream_ << buffer_[i] << "\n";
|
||||
}
|
||||
current_pos_ = 0;
|
||||
}
|
||||
|
||||
FileWriter::FileWriter(const std::string& filename, size_t buffer_size)
|
||||
: buffer_(buffer_size)
|
||||
, current_pos_(0)
|
||||
, total_writes_(0)
|
||||
{
|
||||
std::filesystem::path file_path(filename);
|
||||
std::filesystem::path dir_path = file_path.parent_path();
|
||||
std::filesystem::create_directories(dir_path);
|
||||
stream_.open(filename);
|
||||
}
|
||||
|
||||
FileWriter::~FileWriter()
|
||||
{
|
||||
write_buffer();
|
||||
}
|
||||
|
||||
void FileWriter::write(const Record& record)
|
||||
{
|
||||
if (current_pos_ >= buffer_.size()) {
|
||||
write_buffer();
|
||||
}
|
||||
buffer_[current_pos_] = record;
|
||||
current_pos_++;
|
||||
total_writes_++;
|
||||
}
|
||||
|
||||
void FileWriter::write_chunk(std::span<const Record>& chunk)
|
||||
{
|
||||
for (const auto& record : chunk) {
|
||||
write(record);
|
||||
}
|
||||
}
|
||||
|
||||
size_t FileWriter::total_writes() const
|
||||
{
|
||||
return total_writes_;
|
||||
}
|
||||
67
src/main.cpp
67
src/main.cpp
|
|
@ -1,19 +1,16 @@
|
|||
#include "config.hpp"
|
||||
#include "record.hpp"
|
||||
#include "sort.cpp"
|
||||
#include "writer.hpp"
|
||||
#include <fstream>
|
||||
#include "../include/config.hpp"
|
||||
#include "../include/file_reader.hpp"
|
||||
#include "../include/file_writer.hpp"
|
||||
#include "../include/merge_sorter.hpp"
|
||||
#include "../include/record.hpp"
|
||||
#include <iostream>
|
||||
#include <string>
|
||||
#include <sys/stat.h>
|
||||
#include <sys/types.h>
|
||||
|
||||
void generate_file(int N, const std::string& filename)
|
||||
{
|
||||
std::filesystem::path dirPath = std::filesystem::path(filename).parent_path();
|
||||
std::filesystem::create_directories(dirPath);
|
||||
|
||||
std::ofstream out_stream(filename);
|
||||
Writer output(out_stream);
|
||||
FileWriter output(filename);
|
||||
|
||||
for (int i = 0; i < N; i++) {
|
||||
Record rec = Record::random();
|
||||
|
|
@ -21,47 +18,33 @@ void generate_file(int N, const std::string& filename)
|
|||
}
|
||||
}
|
||||
|
||||
void read_and_evaluate(std::string filename, int n)
|
||||
void read_and_evaluate(std::string filename)
|
||||
{
|
||||
Buffer main_buffer(n);
|
||||
auto buff = main_buffer.divide(1)[0];
|
||||
|
||||
std::ifstream in_stream(filename);
|
||||
Reader input_reader(in_stream);
|
||||
|
||||
FileReader input(filename);
|
||||
while (true) {
|
||||
size_t records_read = read_chunk(input_reader, buff);
|
||||
|
||||
if (records_read == 0) {
|
||||
break;
|
||||
auto rec = input.read();
|
||||
if (!rec.has_value()) {
|
||||
return;
|
||||
}
|
||||
|
||||
for (size_t i = 0; i < records_read; ++i) {
|
||||
const Record& rec = buff[i];
|
||||
std::cout << "[Evaluate i = " << i << ": " << rec.evaluate() << "] " << rec << "\n";
|
||||
}
|
||||
std::cout << "[Evaluate: " << rec.value().evaluate() << "] " << rec.value() << "\n";
|
||||
}
|
||||
}
|
||||
|
||||
int main(int argc, char** argv)
|
||||
{
|
||||
try {
|
||||
Configuration opts = Configuration::parse_args(argc, argv);
|
||||
Configuration opts(argc, argv);
|
||||
|
||||
if (opts.generate_data) {
|
||||
generate_file(opts.N, opts.output_file);
|
||||
return 0;
|
||||
}
|
||||
|
||||
if (opts.evaluate_file) {
|
||||
read_and_evaluate(opts.input_file, opts.N);
|
||||
return 0;
|
||||
}
|
||||
|
||||
sort_file(opts);
|
||||
|
||||
} catch (const std::exception& e) {
|
||||
std::cerr << "Error: " << e.what() << '\n';
|
||||
return 1;
|
||||
if (opts.generate_data) {
|
||||
generate_file(opts.N, opts.output_file);
|
||||
return 0;
|
||||
}
|
||||
|
||||
if (opts.evaluate_file) {
|
||||
read_and_evaluate(opts.input_file);
|
||||
return 0;
|
||||
}
|
||||
|
||||
MergeSorter sorter(opts.b, opts.n, opts.tmp_dir);
|
||||
sorter.sort_file(opts.input_file, opts.output_file);
|
||||
}
|
||||
|
|
|
|||
46
src/merge_sorter.cpp
Normal file
46
src/merge_sorter.cpp
Normal file
|
|
@ -0,0 +1,46 @@
|
|||
#include "../include/merge_sorter.hpp"
|
||||
|
||||
size_t MergeSorter::create_initial_runs(const std::string& input_file)
|
||||
{
|
||||
FileReader reader(input_file);
|
||||
std::span<Record> chunk(buffer_);
|
||||
size_t run_count = 0;
|
||||
|
||||
for (run_count = 0;; run_count++) {
|
||||
size_t records_read = reader.read_chunk(chunk);
|
||||
if (records_read == 0) {
|
||||
break;
|
||||
}
|
||||
|
||||
std::sort(chunk.begin(), chunk.begin() + records_read);
|
||||
|
||||
std::string run_file = tmp_dir_ + "pass0/run_" + std::to_string(run_count) + ".dat";
|
||||
FileWriter writer(run_file);
|
||||
auto output = std::span<const Record>(chunk.data(), records_read);
|
||||
writer.write_chunk(output);
|
||||
|
||||
disk_reads_ += reader.total_reads() - disk_reads_;
|
||||
disk_writes_ += writer.total_writes() - disk_writes_;
|
||||
}
|
||||
|
||||
return run_count;
|
||||
}
|
||||
MergeSorter::MergeSorter(size_t buffer_rows, size_t buffer_cols, const std::string& tmp_dir)
|
||||
: buffer_(buffer_rows * buffer_cols)
|
||||
, tmp_dir_(tmp_dir)
|
||||
, buffer_rows(buffer_rows)
|
||||
, disk_reads_(0)
|
||||
, disk_writes_(0)
|
||||
, phases_(0)
|
||||
{
|
||||
if (buffer_rows < 3) {
|
||||
throw std::invalid_argument("Buffer must have at least 3 rows for merging");
|
||||
}
|
||||
}
|
||||
|
||||
bool MergeSorter::sort_file(const std::string& input_file, const std::string& output_file)
|
||||
{
|
||||
create_initial_runs(input_file);
|
||||
std::cout << output_file << '\n';
|
||||
return true;
|
||||
}
|
||||
46
src/record.cpp
Normal file
46
src/record.cpp
Normal file
|
|
@ -0,0 +1,46 @@
|
|||
#include "../include/record.hpp"
|
||||
#include <random>
|
||||
|
||||
int Record::evaluate() const
|
||||
{
|
||||
int out = 0;
|
||||
int x_n = 1;
|
||||
for (int i = 0; i < 5; i++) {
|
||||
out += a[i] * x_n;
|
||||
x_n *= x;
|
||||
}
|
||||
return out;
|
||||
}
|
||||
|
||||
bool Record::operator<(const Record& other) const
|
||||
{
|
||||
return evaluate() < other.evaluate();
|
||||
}
|
||||
|
||||
Record Record::random()
|
||||
{
|
||||
static std::random_device rd;
|
||||
static std::mt19937 gen(rd());
|
||||
static std::uniform_int_distribution<int> dist;
|
||||
|
||||
Record r {};
|
||||
for (int& val : r.a) {
|
||||
val = dist(gen);
|
||||
}
|
||||
r.x = dist(gen);
|
||||
return r;
|
||||
}
|
||||
|
||||
std::ostream& operator<<(std::ostream& os, const Record& r)
|
||||
{
|
||||
|
||||
os << r.a[0] << ' ' << r.a[1] << ' ' << r.a[2] << ' '
|
||||
<< r.a[3] << ' ' << r.a[4] << ' ' << r.x;
|
||||
return os;
|
||||
}
|
||||
|
||||
std::istream& operator>>(std::istream& is, Record& r)
|
||||
{
|
||||
is >> r.a[0] >> r.a[1] >> r.a[2] >> r.a[3] >> r.a[4] >> r.x;
|
||||
return is;
|
||||
}
|
||||
Loading…
Add table
Add a link
Reference in a new issue