Memory and the Storage Hierarchy
The storage hierarchy balances access speed, capacity, and cost by keeping frequently used data near the processor.
Summary
The storage hierarchy balances access speed, capacity, and cost by keeping frequently used data near the processor. The usual order from fastest and smallest to slowest and largest is registers, L1/L2/L3 caches, DRAM, SSDs, and HDDs.
| Storage level | Main role | Key property |
|---|---|---|
| Registers and caches | Hold immediately needed or frequently accessed data | Small and fast |
| DRAM main memory | Holds active programs and data | Larger than caches; volatile |
| SSD and HDD storage | Stores persistent data and can back displaced pages | Larger but generally slower |
Temporal locality means recently accessed data or instructions are likely to be reused. Spatial locality means nearby addresses are likely to be accessed soon. Caches transfer fixed-size cache lines, so contiguous data and predictable access patterns often improve performance. Conventional SRAM caches and DRAM lose their contents without power.
For a single cache level, average memory access time is , where is hit time, is the miss rate as a fraction, and is the additional miss delay.
🧠 Key Concepts
- Storage hierarchy
- Temporal locality
- Spatial locality
- Cache lines
- Memory volatility
- Average memory access time
- Virtual memory
- Latency versus bandwidth
- Page faults
🧠 Quick Check
See what you remember from the summary.
A program traverses contiguous array elements in order. Which locality principle explains why fetching a cache line can help subsequent accesses?
🧠 Flashcards Preview
Tap a card to reveal the definition.
Ready to quiz yourself?
Test what you remember with a full practice quiz on this note. Create a free account and start in seconds.
Full Notes
Read the original note content before deciding whether to save or study from it.
Memory and the Storage Hierarchy
📘 Overview A computer's storage hierarchy balances access speed, capacity, and cost by placing frequently used data near the processor. Registers, caches, main memory, and persistent storage provide progressively larger but generally slower storage.
🧠 Key Idea The hierarchy exploits locality: programs tend to reuse recently accessed data and access nearby locations. Small, fast storage holds the active working set, reducing how often the processor must access slower levels.
⚔️ Core Details: - The usual order from fastest and smallest to slowest and largest is processor registers, L1/L2/L3 caches, DRAM main memory, SSD storage, and HDD storage. - Conventional SRAM caches and DRAM main memory are volatile: their contents are lost when power is removed. - Caches transfer data in fixed-size cache lines rather than fetching only the individual byte requested. - Virtual memory maps virtual addresses to physical memory, with secondary storage available as backing for pages displaced from RAM. - For a single cache level, average memory access time is
More ways to study when you copy this note
Copy this note into your library to unlock focused practice sessions and long-term review.
Answer all questions first, then see feedback at the end — the way real exams work.
Focuses each session on what you got wrong, not what you already know.
Full timed exam with all questions, no pausing, and results at the end. Built for board exam prep.
Part of 🗄️ BS Information Systems — 1st Year
Browse this Review Set →More Computer Science notes
See all →More in Computer Science
See all →More from NoteLib
Browse NoteLib's public notes →Copy this note to your library and get the full Study Pack instantly — summary, key concepts, and practice quiz included.