Workshop: Advanced Java Profiling

Location: Rutherford, 4th flr.

Duration: 9:00am - 4:00pm

Day of week: Friday

Level: Intermediate

Key Takeaways

  • Profilers - the good, the bad, the ugly?
  • Sampling vs recording
  • Sampling bias
  • Flamegraphs and other visualisations
  • Allocation profiling
  • Benchmarking
  • Custom tracepoints
  • Profiling in production
  • OS/Kernel/Hardware profiling

Prerequisites

Attendees will need a laptop with at least 4 CPUs.

The workshop material will be run from within a virtual machine, so attendees should install the latest versions of:

Users with a Linux laptop will be able to run the project natively, but openjdk-11 and openjdk-8 should be installed.

More talks on:

Profiling is a key technique in observing and understanding application performance. This ability allows us to determine where in our systems bottlenecks lie, and gives us the information required to target those bottlenecks. Armed with this knowledge and the use of appropriate design patterns, we can fully utilise the power of modern computing hardware, making our applications faster. Aiming for good application performance has many benefits: better user experience, superior scalability, and improved efficiency, which can be particularly useful in today’s cloud deployments.   

This course is designed for application developers and performance engineers, who are already well-versed in the use of standard application profilers. The material will build on this knowledge, exploring in more depth different approaches to profiling, and covering more advanced techniques. Other topics include measuring and reporting system latency, and profiling in production.

Speaker: Mark Price

Performance Engineering Specialist at Aitu Software

Mark Price is an experienced technologist, with an enthusiasm for all things performance-related. He writes a popular blog on JVM and Linux topics, and is a regular speaker at international tech conferences. 

Mark has worked on high-throughput, large-scale, and low-latency distributed systems for more than ten years. He has also authored and contributed to many open-source projects that are used for monitoring, messaging and profiling high-performance systems. 

When not working as a consultant in Performance Engineering, Mark is interested in exploring the limits of what can be achieved with fast computers, mechanical sympathy, and an understanding of how our programs interact with the operating-system that they run on.

Find Mark Price at

Speaker page

https://epickrram.blogspot.com/
@epickrram
https://uk.linkedin.com/pub/mark-price/1/7b4/645
GitHub

Tracks

Schedule

Follow QCon

Contact

Menu

QCons around the World

This site is protected by reCAPTCHA and the Google Privacy Policy and Terms of Service apply.