Computer Systems Fundamentals

Lecture Plan Overview

• Topics include bits and context, data vs. information, program translation, execution, , caching & storage, & the OS role.

System Understanding

• Need to understand systems to program them effectively.
• All computers share the same basic hardware like CPUs/processors, RAM, memory disks, & storage.
• Software includes operating systems, programs, system utilities, & applications.
• Components are interdependent; a missing/malfunctioning component renders a computer useless.
• Focus on software development necessitates understanding all underlying elements.

Core Concepts

• Programs manipulate information.
• Information comprises bits + context
• Translation reconciles program & hardware requirements
• Programs may use large numbers, but hardware typically uses 32-bit operations
• A single high-level program instruction can translate into multiple machine instructions.
• Program execution depends on the hardware architecture.
• They are affected by the OS, memory architecture, and caching mechanisms.

Bits and Context

• Data alone lacks inherent meaning.
• Data derives meaning from its interpretation or context.
• Data with the same meaning can have multiple representations such as ASCII, Hex & Binary.

Data vs. Information

• Data is raw symbols, while information is structured/interpreted data providing meaning.
• The example code snippet is information, while its hexadecimal representation is data.

Data Meanings

• Same data presents different meanings based on the context
• The stream of consciousness looks different depending on which interpreter translates it:
  • Text
  • 8-bit integers
  • 32 bit integers

Data Representations

• The data can be represented using different methods
  • ASCII Bytes
  • Hexadecimal Bytes
  • Decimal Bytes

Context Importance

• Context is essential for data to be meaningful
• The "hello.c" data can be interpreted as text, 8-bit integers, or 32-bit integers, etc.
• Everything from files and memory to the CPU relies on interpretation.
• System manipulation requires understanding data representation and interpretation.

Program Translation

• Program translation is fundamental to system operation.
• Programs undergo multiple translations to run on a system.
• Java program (.java) translates to bytecode (.class) and then to machine code, running on the JVM.
• C program (.c) translates to assembly (.s), then to object code (.o), and finally to an executable.
• Programs undergo transformation at each translation stage.
• Each translation phase relies on the previous stage and the translator.
• Every step has an affect on the final output of a compilation.

C Compilation

• Includes preprocessing, compilation, assembly & linking to create an executable from C code.
• The command gcc -Og -S hello.c generates the assembly code "hello.s"
• The command gcc -Og -c hello.c creates the object file "hello.o"
• The command gcc -Og -o hello.o hello.c generates the executable "hello"
• Use ./hello command to run the code

Assembly File

• Is text based code that relates directly to the functional operation of the computer
• Compiled using GCC to create a computer program

Program Performance

• Efficient code writing is important
• Code choices impact performance such as: switch vs if statements
• Loop selection matters: for, while, and do-while
• Choose between references and indexing

Avoiding Security Holes

• Compilation process insight prevents vulnerabilities, like stack/buffer overflows, and uninitialized memory
• Understanding implementation defends against vulnerabilities

Linking Errors

• Resolved by understanding multiple global variable/function definitions and header files
• Correct linking order resolves errors

Program Execution

• Program execution depends on hardware.
• Consider the CPU and its instruction set, word size, and implementation
• RAM considerations include the available memory, speed, and caching.
• Bus factors involve word size, speed, and coherency protocols.

Hardware Overview

• The hardware is made up of the CPU, ALU, RAM, and a programmable interrupt controller

Execution Process

• CPU executes code via fetch-decode-execute loop:
• Fetch instructions from memory (location in Program Counter- PC )
• Decode to understand the operation
• Execute based on the operation and instruction given
• Instructions may read/write to memory and modify the CPU state
• The program may jump around, which will modify PC

Instruction Speed

• Memory access instructions are slower due to data transfer across the bus.
• Memory is generally slower than cache.
• Programs made faster by reducing memory accesses and optimizing cache use.
• Computation can be sped up by speeding up memory access

Caching

• Smaller memory sits between the large memory and the CPU, storing common data
• Improves data access speed
• Components closer to the CPU
• More expensive

Memory Hierarchy

• Faster and small memory sits between the processor and slower, larger storage
• Each level acts as a cache for the level below it.
• Key considerations include data location, usage time, and caching strategy.
• Focus on the memory hierarchy and cache management.

OS Role

• The OS manages the computer hardware including CPU, I/O, & RAM.
• Provides an execution environment for programs.
• Acts as an intermediary between the computer, users, and programs.
• Protects hardware, programs, and users from each other.
• Computer resources are accessed through the OS.

System Resources

• CPU abstraction is a process.
• RAM abstraction is a memory address space.
• Secondary storage i.e. hard drive abstraction is a file.
• The abstraction for Network is a file.
• Abstraction for Peripheral Devices is a file.

Key Abstraction: Process

• It's an instance of a running program.
• Is associated with a memory space
• and may have files open for access

Key Abstraction: Memory

• The range of memory locations that a process can access.
• Locations are uniquely accessible via memory addresses
• Each location is read, written to, or executed

Key Abstraction: Files

• A file is a stream of data identified via a name in the OS;s file system or a file descriptor

Threads

• Processes can run concurrently.
• Traditional systems are sequential but concurrent.
• Instructions of different Processes may be interleaved.
• Context switching is important.

Virtual Memory

• It gives each running process exclusive use of memory, with specific memory areas
• Includes kernel virtual memory, user stack, shared libraries, and more.

Bytes

• Files are a byte sequence.
• I/O devices are modeled as files
• Applications gain a uniform I/O view.

Process Composition

• Composed of CPU context, memory space, and execution threads.
• Memory involves a stack, heap, static variables, and running logic.

Execution Thread

• Active execution of a program's code at a specific stored PC location
• Depends on the CPU context & memory.

Key Takeaways

• Information equates data + its context
• Data is represented & interpreted differently.
• Programs need translation before execution.
• Translation affects execution.
• Operating System, hardware and architecture are important
• Memory access dominates execution cost.
• System design mitigates access costs.
• The OS manages the system providing execution mechanisms, abstractions, and environments.

Exploring fundamental concepts of computer systems. Covering data representation, program translation, the role of operating systems, and storage hierarchies. Includes essential abstractions and components.