Digital Design & Computer Architecture D13: Problem-Solving Session 13 (Spring 2026)

Item: Digital Design & Computer Architecture D13: Problem-Solving Session 13 (Spring 2026)
Rating: 4
Author: PostSingularity Institute

by Onur Mutlu

Formal & Physical Sciences Computing & Cybersecurity UYComputer science UYFComputer architecture and logic design

🎙 Onur Mutlu 👥 64K 📅 June 1, 2026 ⏱ 28 min 👁 895 📄 tutorial 🧭 2026-08-15

Available in: English (current) Français

Keywords

prefetchingcachecoverageaccuracybandwidth

Summary

This video is the 13th problem-solving session for the Digital Design and Computer Architecture course at ETH Zürich, taught by Professor Onur Mutlu. The session focuses on prefetching in computer architecture. The instructor begins with a brief review of prefetching concepts, including its goal to reduce access latency and eliminate compulsory misses, and introduces key metrics: coverage, accuracy, and timeliness. He then explains two common prefetchers: the stride prefetcher and the stream prefetcher. The main part of the session is dedicated to solving a set of exercises. The first exercise involves analyzing a simple stride prefetcher on a given memory access pattern, calculating its coverage and accuracy (both 0%). The second exercise introduces a next-n-block prefetcher, and the instructor guides through finding the value of n that yields a coverage of 66.67% and accuracy of 50%, concluding n=2. He then calculates the bandwidth overhead for this prefetcher as 5/3. The final exercise asks for the minimum n to achieve 100% coverage in steady state, which is n=5, and the corresponding bandwidth overhead is 7/3. The session concludes with a brief Q&A and well-wishes for the exam.

186 words

Critical Evaluation

Value of the Information & Strength of the Argument

The value of the information is high for students learning about prefetching, as it provides a hands-on, problem-solving approach that reinforces theoretical concepts. The argumentation is solid: the instructor systematically walks through each problem, clearly defining metrics and showing step-by-step calculations. He also addresses potential pitfalls, such as the importance of steady-state analysis and the handling of duplicate requests via MSHRs. The reasoning is logical and easy to follow, making complex topics accessible.

Scientific Rigor, Source Quality, Title Accuracy

The scientific rigor is high, as the content is technically accurate and presented by an expert in the field. The session is part of a university course, and the instructor demonstrates a deep understanding of the subject. The description includes several relevant academic papers on memory-centric computing and RowHammer, which provide additional context and sources for further study. The title accurately reflects the content, and the video is well-structured. No comments were provided for analysis.

163 words

Title / Content Match

The title accurately reflects the content: a problem-solving session for a digital design and computer architecture course.

Quality & Reliability

8/10

The lecture is a well-structured problem-solving session by a recognized expert in computer architecture. The content is technically accurate, and the reasoning is clear and methodical. The session is part of a university course, and the instructor demonstrates deep knowledge. However, it is a tutorial rather than a peer-reviewed source, and the video does not provide citations for the specific claims made, though the description includes relevant academic references.

Key Moments

Markers derived by PSI from the transcript: the creator did not define chapters.

Cited Sources

A Modern Primer on Processing in Memory — Recommended reading for the course, related to memory-centric computing.
Memory-Centric Computing: Solving Computing's Memory Problem — Recommended reading, discusses memory-centric computing.
Memory-Centric Computing: Recent Advances in Processing-in-DRAM — Recommended reading, recent advances in processing-in-DRAM.
Intelligent Architectures for Intelligent Computing Systems — Recommended reading, intelligent architectures.
RowHammer: A Retrospective — Recommended reading, RowHammer retrospective.
Fundamentally Understanding and Solving RowHammer — Recommended reading, understanding and solving RowHammer.
Accelerating Genome Analysis via Algorithm-Architecture Co-Design — Recommended reading, genome analysis acceleration.
From Molecules to Genomic Variations: Accelerating Genome Analysis via Intelligent Algorithms and Architectures — Recommended reading, intelligent genome analysis.

Concurring Sources

A Modern Primer on Processing in Memory — Provides background on memory-centric computing, which is related to prefetching.
Memory-Centric Computing: Solving Computing's Memory Problem — Discusses memory-centric computing, relevant to prefetching.

External References

Contribution & Novelties

This session provides a practical, problem-solving approach to understanding prefetching, which is often taught only theoretically. It offers clear examples of how to calculate coverage, accuracy, and bandwidth, and highlights the importance of steady-state analysis. The session also clarifies the behavior of different prefetcher designs on a specific access pattern, which is valuable for students.

Pour aller plus loin :

Cache prefetching - Wikipedia — Overview of prefetching techniques and metrics.
Miss Status Holding Register (MSHR) - Wikipedia — Explanation of MSHRs, which are mentioned in the session.
Stride prefetcher - Wikipedia — Detailed description of stride prefetching.

98 words

Radar Profile

The radar profile shows high scores in quality of information, technical level, and reliability, with a slightly lower score for quantity of information. This indicates a focused, in-depth tutorial that excels in accuracy and depth but may not cover a broad range of topics.

Reliability 8/10