UTPA STEM/CBI Courses/Microprocessor Systems/Microprocessor System Design

From Wikiversity
Jump to navigation Jump to search

Course Title: Microprocessor System Design

Lecture Topic: Cache Memory Design

Instructor: Dr. Yul Chu

Institution: University of Texas - Pan American

Backwards Design[edit | edit source]

Course Objectives

  • Primary Objectives- By the next class period students will be able to:
    • Understanding memory hierarchy.
      Memory Hierarchy
    • Understand how the mapping function works.
    • Understand how the replacement policy works.
    • Understand how the writing policy works.
Cache Memory Access
  • Sub Objectives- The objectives will require that students be able to:
    • Understand how the read memory operation works.
    • Understand how the write memory operation works.
  • Difficulties- Students may have difficulty:
    • To understand how to access cache memory with memory address.
    • To understand how to replace the block data from memory to cache memory.
Memory locations vs. Cache Memory Locations
  • Real-World Contexts- There are many ways that students can use this material in the real-world, such as:
    • Intel x86 CISC microprocessor
    • MIPS 2000 RISC processor

Model of Knowledge

  • Concept Map
    • Determining a cache size
    • Determining a block size
    • Understanding mapping functions
    • Understanding replacement policies
    • Understanding write policies
  • Content Priorities
    • Enduring Understanding
      • Direct-mapped, n-way set-associative, and fully associative
      • Least Recently Used (LRU), Most Recently Used (MRU), and Random replacement policies
      • Write-Back/Write-Allocate and Write-Through/Write-No-Allocate policies
    • Important to Do and Know
      • How to access cache memory with index part of memory address.
      • How to replace a cache line based on dirty bit info.
      • The concept of block data and byte offset is ?
    • Worth Being Familiar with
      • Understanding cache memory and memory locations
      • Computing average memory access time
      • Understanding block address and data in cache memory

Assessment of Learning

  • Formative Assessment
    • In Class (groups)
      • Let each group (3 to 4 students/group) design a cache memory and discuss it.
      • Let each group check the five factors for a cache memory and verify it.
    • Homework (individual)
      • To implement some memory references manually.
      • To compute average memory access time for direct-mapped and n-way set-associative cache memory.
  • Summative Assessment
    • Let each group design a cache memory (4-way set-associative) and implement several read and write operations.
    • Final Exam

Legacy Cycle[edit | edit source]

OBJECTIVE

By the next class period, students will be able to:

  • Understand how the read memory operation works
  • Understand how the write memory operation works.
  • Understand why the temporal and spatial localities are important to design a cache memory.

The objectives will require that students be able to:

  • Understand memory hierarchy.
  • Understand physical memory and virtual memory.
    The program thinks it has a large range of contiguous addresses; but in reality the parts it is currently using are scattered around memory (RAM), and the inactive parts are saved in a disk file.
  • Understand how the mapping function works.
  • Understand how the replacement policy works.
  • Understand how the writing policy works.

THE CHALLENGE

How to design cache memory to reduce cache misses for a small L1 cache to improve system performance?

GENERATE IDEAS

How to design a cache memory using localities?

MULTIPLE PERSPECTIVES

How to implement 'write hit and miss policies'? How to implement 'LRU replacement policy for 2-way or 4-way set-associative cache memories'? How to update the cache memory for a write-back? How to access a cache memory from processor?

RESEARCH & REVISE

-Read lecture notes, textbook, and papers. -group discussion. -do manual implementation and revise any errors.

TEST YOUR METTLE

  • Formative assessment: Small quizzes, concept quizzes, and discussion.
  • Summative Assessment: Projects and exam.

GO PUBLIC

Final presentation, report, grading, and verification test.

Pre-Lesson Quiz[edit | edit source]

  1. What is the purpose of cache memory?
  2. Is it possible to access cache memory with memory address? If yes, how?
  3. What does SRAM stand for?
  4. What does RAM stand for?
  5. Which one is more expensive, RAM or SRAM ?
  6. What is L1 (Level One) cache?
  7. What is Temporal Locality?
  8. Explain Fetch-Execute cycle.
  9. Explain the differences between RISC and CISC.
  10. What is Spatial Locality?


Test Your Mettle Quiz[edit | edit source]

  1. Why the performance of 2-way set-associative has better performance than direct-mapped cache?
  2. Which of the following replacement policies is popular and why? 1) LRU 2)MRU 3) Random
  3. For a small cache memory, which of the following cache misses is critical to performance? 1) Conflict miss 2) Compulsory miss 3) Capacity miss
  4. What kind of write-miss policy can work with write-back hit policy?
  5. List 5 factors to design cache memory.
  6. List three steps to determine cache miss or hit?
  7. What is memory hierarchy?
  8. Explain physical address and virtual address in brief.
  9. What are the benefits of the virtual address?
  10. If the size of a direct-mapped cache memory is 8KB and the block size is 64 bytes, then how many bits for cache indexing are needed?