3,079 views
The nearest conference is JPoint 2025, April 3–4 (Moscow + broadcast). Details and tickets: https://jrg.su/T2zfbS — — . . . . The report is devoted to the peculiarities of profiling small sections of an application that greatly affect the characteristics of the entire application. What if the profiled code is executed for 0.0001 seconds, but you want to speed it up to 0.00005 seconds? We will consider the limits of Java profilers and how to squeeze the most out of the standard Linux profiler — perf. Next, we will look at what hardware features of processors will help us get a representative profile even faster. At the end, we will talk about Intel Processor Trace technology, which allows you to take another step in profiling accuracy and reconstruct the execution of a program section.