Free courses from frontend to fullstack and AI
AI, Data Science & Cloud Certificates from Google, IBM & Meta
Overview
Google, IBM & Meta Certificates – 40% Off
One plan covers every Professional Certificate on Coursera.
Unlock All Certificates
Explore a 17-minute conference talk from USENIX OSDI '24 that introduces a novel sampling-based profiling technique called "blocked samples" for comprehensive analysis of both on-CPU and off-CPU events in modern computer systems. Learn about two new profilers, bperf and BCOZ, which leverage this technique to provide detailed symbol-level profile information and perform causality analysis to identify performance bottlenecks accurately. Discover how this approach addresses the limitations of conventional profiling methods in assessing overhead across various bottlenecks in contemporary applications. Gain insights into the verification process using real applications and understand how the optimization results align with BCOZ's predicted performance improvements through causality analysis.
Syllabus
OSDI '24 - Identifying On-/Off-CPU Bottlenecks Together with Blocked Samples
Taught by
USENIX