Optimization-Aware Compiler-Level Event Profiling
Offered By: ACM SIGPLAN via YouTube
Course Description
Overview
Explore a novel technique for quantifying compiler-internal events in optimized code, reducing profiling perturbation on compiler optimizations. Delve into the implementation strategies based on path-profiling and a modification to the standard path-profiling algorithm for use in modern just-in-time (JIT) compilers. Examine the application of this technique in analyzing optimization behaviors in Graal, a state-of-the-art compiler for the Java Virtual Machine. Learn how this approach identifies reasons behind performance improvements and unexpected slowdowns in specific optimizations. Discover the significantly lower execution-time overhead achieved by the proposed implementations compared to naive implementation methods. This 19-minute conference talk from ACM SIGPLAN's OOPSLA23 presents groundbreaking research in compiler-level event profiling, offering valuable insights for developers and researchers in the field of dynamic analysis and compiler optimization.
Syllabus
[OOPSLA23] Optimization-Aware Compiler-Level Event Profiling
Taught by
ACM SIGPLAN
Related Courses
Cadenza - Building Fast Functional Languages FastGOTO Conferences via YouTube Performance Tuning Twitter Services with Graal and Machine Learning
Devoxx via YouTube Performance Tuning Twitter Services with Graal and Machine Learning
Devoxx via YouTube Twitter's Quest for a Wholly Graal Runtime
Devoxx via YouTube Performance Tuning Twitter Services with Graal and Machine Learning
Devoxx via YouTube