CPE 4520 - Memory System and I/O System Architecture

3 lecture hours 0 lab hours 3 credits
Course Description
High-performance computing relies on an efficiently designed memory and I/O pyramid to move information between mass storage and the processor at a rate sufficient to meet computing demands. This course explores memory and I/O systems architecture, components, and communication protocols through classroom reading, case studies, and homework exercises demonstrating mastery of terminology, concepts, logic design, simulation, and calculations.
Prereq: CPE 1510  (quarter system prereq: CE 1921 or CS 2711)
Note: None
This course meets the following Raider Core CLO Requirement: None
Course Learning Outcomes
Upon successful completion of this course, the student will be able to:
  • Diagram the memory and I/O pyramid with associated technologies identified
  • Articulate primary and secondary memory characteristics
  • Describe multi-level cache organization, replacement algorithms, and coherence algorithms
  • Describe virtual memory held in secondary storage and its effect on performance
  • Describe the use of direct memory access and its effect on performance
  • Describe the logical structure and access behavior of SRAM and DRAM
  • Describe the role of controller hubs in system architecture
  • Describe the port structure, signals, and protocols of the ISA, PCI, PS/2, and USB busses
  • Evaluate the performance of a CPU with memory and I/O under multiple realistic constraints

Prerequisites by Topic
  • Assembly language programming
  • Design and simulation of scalar pipelines using a hardware description language
  • Performance evaluation of central processing units (CPUs) under realistic constraints

Course Topics
  • Memory and I/O as stored program and data repositories
  • Speed and size placement on the memory and I/O pyramid  
  • Multi-level on-chip cache RAM to reduce pipeline stalls caused by memory latency
  • Compulsory, capacity, and conflict cache misses
  • Direct, set-associative, and fully associative cache organization
  • Write-through and write-back data replacement algorithms
  • Cache sizing, replacement algorithm, and organization effects on cache miss rate
  • Virtual memory, virtual and physical memory addresses, and paging
  • Page tables, translation lookaside buffers, and virtual memory protection
  • Direct memory access (DMA)
  • Digital logic structure of static RAM (SRAM) and dynamic RAM (DRAM)
  • Transistor structure of SRAM (six transistor memory cell, poly-load memory cell) and DRAM
  • Row address and column address strobes, sense amplifiers, and restorative logic
  • Logical waveforms of memory read and write cycles
  • Dynamic RAM refresh cycles
  • Double data rate memory (DDR)
  • Memory and I/O controller hubs (north and south bridges)
  • Frontside and backside busses
  • Programmed and interrupt-driven I/O models
  • Intel IBM PC System architecture
  • Basic Input/Output System (BIOS) firmware and booting process
  • Master Boot Record
  • Plug-and-play technology
  • Advanced Configuration and Power Interface (ACPI)
  • Unified Extensible Firmware Interface (UEFI) as a BIOS replacement
  • Parallel and serial communication busses
  • Industry Standard Architecture (ISA) parallel bus socket, signals, and protocol waveforms
  • Peripheral Component Interconnect (PCI) parallel bus socket, signals, and protocol waveforms
  • PS/2 serial bus port, signals, and protocol waveforms
  • Universal Serial Bus (USB) ports, signals, and protocol waveforms
  • Performance evaluation of CPUs with memory and I/O systems under multiple realistic constraints

Coordinator
Dr. Russ Meier


Print-Friendly Page (opens a new window)