Skip to content

Latest commit

 

History

History
95 lines (68 loc) · 4.35 KB

README.md

File metadata and controls

95 lines (68 loc) · 4.35 KB

Virtual Memory Concepts

Overview

This directory contains documentation about fundamental virtual memory concepts that apply across different operating systems. Virtual memory is a memory management technique that provides an abstraction layer between applications and physical memory, enabling efficient memory usage, protection, and sharing.

Available Documentation

The following PDF resources are available in this directory:

  • ch9.pdf: Comprehensive chapter on virtual memory fundamentals
  • CSE351-L02-memory-I_17wi.pdf: University course materials on memory basics
  • Lecture21.pdf: Academic lecture on virtual memory implementation
  • Lecture_2_Addressing_Data_in_Memory_Microprcessor-Design.pdf: Memory addressing in modern processors
  • Operating System Virtual Memory.pdf: OS-level virtual memory implementation
  • os_virtual_memory.pdf: Alternative perspective on OS virtual memory
  • publication_3_22584_1575.pdf: Research publication on virtual memory optimization

Core Virtual Memory Concepts

Memory Address Translation

  • Virtual Address: Memory address as seen by processes
  • Physical Address: Actual hardware memory location
  • Translation Lookaside Buffer (TLB): Cache for address translations
  • Page Tables: Data structures mapping virtual addresses to physical addresses

Paging

  • Page: Fixed-size block of memory (typically 4KB to 64KB)
  • Page Frame: Physical memory unit corresponding to a page
  • Page Fault: Exception raised when accessed memory isn't in physical RAM
  • Demand Paging: Loading pages into memory only when needed

Memory Protection

  • Protection Bits: Read/Write/Execute permissions
  • Rings/Privilege Levels: Separation between user/kernel modes
  • Memory Isolation: Preventing processes from accessing each other's memory
  • Shared Memory: Controlled sharing of memory between processes

Page Replacement Algorithms

  • Least Recently Used (LRU): Removes pages that haven't been used for longest time
  • First-In-First-Out (FIFO): Removes oldest pages first
  • Clock Algorithm: Approximation of LRU with lower overhead
  • Working Set Model: Tracks active memory pages a process needs

GPU Virtual Memory

CUDA Virtual Memory Management

  • Virtual Address Reservation: Pre-allocating virtual address space for GPU operations
  • Physical Memory Allocation: Allocating GPU memory with specific alignment requirements
  • Memory Mapping: Connecting virtual addresses to physical GPU memory
  • Access Control: Setting permissions for GPU memory regions
  • Memory Release: Properly deallocating GPU memory resources

Additional Resources

Online Resources

Books

  • "Operating Systems: Three Easy Pieces" - Remzi H. Arpaci-Dusseau and Andrea C. Arpaci-Dusseau
  • "Modern Operating Systems" - Andrew S. Tanenbaum
  • "Computer Architecture: A Quantitative Approach" - John L. Hennessy and David A. Patterson
  • "Windows Internals" - Mark E. Russinovich, David A. Solomon, and Alex Ionescu
  • "Understanding the Linux Kernel" - Daniel P. Bovet and Marco Cesati

Key Comparisons

Virtual Memory Across Platforms

Feature Linux Windows macOS
Page Size 4KB default (configurable) 4KB (small) / 2MB (large) 4KB (base) / 16KB (ARM)
Address Space User/Kernel split User/Kernel split User/Kernel split
Swapping Swap partitions/files Page files Swap files
Large Pages Huge pages Large pages Superpage support
Memory Pressure kswapd Working Set Manager Memory pressure handler

Historical Evolution

  • Early systems with no virtual memory
  • Introduction of segmentation
  • Development of paging systems
  • Modern combined approaches
  • Future directions in memory management