Here is the revised README for the ioOptimization project:

ioOptimization

Overview

The ioOptimization project focuses on optimizing disk I/O operations using various C++ programs. The project is divided into multiple parts, each targeting different aspects of I/O optimization such as measurement, caching, system calls, and raw performance.

Setup and Usage

1. Extract the codes.zip file and run the compilation script to compile all provided C++ files.
2. There are three script files:
   • Compilation script: Compiles all programs.
   • Execution script: Executes all compiled code.
   • Extra credit script: Compiles and executes additional code (e.g., dd compare code and Google benchmark).

Scripts

• For the execution and extra credit scripts, you may need to change filenames accordingly.
• If you encounter "Permission denied" issues, run chmod -x <scriptname>.
• For optimal file size execution within a reasonable time, run 'run2' separately and change the filename accordingly in executionScript.sh.
• Default filename in executionScript.sh and extracreditScript.sh is "ubuntu-21.04-desktop-amd64.iso". You can change the filename or run the code manually.

Compilation

• Code compilation command: g++ -std=c++11 -O3 run.cpp -o run

Project Breakdown

Part 1: Basics

Write a program that can read and write a file from disk using the C/C++ library's open, read, write, and close functions.

• Parameters:
  • File name
  • File size (for writing)
  • Block size (how much to read with a single call)
• Execute: ./run <filename> [-r|-w] <block_size> <block_count>

Part 2: Measurement

Write a program to find a file size that can be read in a reasonable time (between 5 and 15 seconds).

• Input: block size
• Output: file size/block count
• Execute: ./run2 <filename> <block_size>

Part 3: Raw Performance

Make your program output the performance achieved in MiB/s and create a graph showing performance as block size changes.

Part 4: Caching

Experiment with clearing the caches and not. Create a graph showing performance for cached and non-cached reads for various block sizes.

• Clear cache: sudo sh -c "/usr/bin/echo 3 > /proc/sys/vm/drop_caches"
• Execute:
  • ./run3 <file_name> block_size
  • ./run3 <file_name> block_size (again for cached performance)

Part 5: System Calls

Measure performance in MiB/s and B/s when using a block size of 1 byte.

• Execute: ./run5a <file_name> 1 and ./run5b <file_name> 1

Part 6: Fast Performance

Optimize your program to run as fast as possible. Report both cached and non-cached performance numbers.

• Execute: ./fast <filename>

Conclusion

This project provides a comprehensive framework for optimizing disk I/O operations. Each part focuses on different optimization techniques, making it a versatile tool for performance analysis and improvement.