Session + Live Q&A
Comparison of Performance of Multiple CPU Architectures
With an increase in the choices of CPU architectures available both on-premise and in the public cloud, it is becoming increasingly important to understand the differences in performance between these options. We will compare performance examples of 4 different CPU vendors while also comparing bare metal and two public cloud enviroments. Our testing will utilize SPECjbb2015 with OpenJDK17 to illustrate various aspects of peak throughput as well as latency response. We'll also consider performance/watt differences measured on bare metal workloads, providing further insight into different architectures. Rather than attempting to declare a "winner", we'll focus on illustrating similiarites and differences among multiple high performing architectures.
Speaker

Matthew Singer
Senior Staff Hardware Engineer @twitter
Matt Singer is Senior Staff Hardware Engineer and architect of Twitter's server platforms. An expert in performance and server architecture, Matt is the technical leader for the hardware team that created one of the most...
Read moreFind Matthew Singer at:
Speaker

Jeff Balk
Sr. Hardware Engineer @twitter
Jeff Balk is a Senior Hardware Performance Engineer at Twitter. He has 15 years experience in the tech industry, back when 65nm and dual core were cutting edge. He has a MSEE from the University of Minnesota.
Read moreFind Jeff Balk at:
From the same track
Unblocked By Design
Monday Nov 8 / 12:10PM EST
Asynchronous designs present an opportunity for greater system performance, but require a different mindset to take full advantage of. Applying normal synchronous approaches to asynchronous designs often lead to less efficient implementations, added complexity, and can negate the advantages a...

Todd Montgomery
Ex-NASA Researcher and High Performance Distributed Systems Whisperer
The Mechanics of Metrics: Aggregation across Dimensions
Monday Nov 8 / 01:10PM EST
The most efficient way to observe application behavior relies on metrics, key/value pairs of numerical data. Capturing and processing metrics can help you identify and respond to potential issues before they escalate and cause more serious problems. But what should you measure? What do those...

Erin Schnabel
Senior Principal Software Engineer @Redhat
JIT vs. AOT: How to Pick the Right Approach
Monday Nov 8 / 02:10PM EST
The popularity of Java introduced the benefits of Just-In-Time (JIT) dynamic compilation to many developers. Recently, Ahead-Of-Time (AOT) compilation techniques have also been available for Java. When should a developer use a JIT or AOT? The panelists will discuss the trade-offs between the two...

Monica Beckwith
Java Champion, First Lego League Coach, passionate about JVM Performance @Microsoft

Gil Tene
CTO @AzulSystems

Todd Montgomery
Ex-NASA Researcher and High Performance Distributed Systems Whisperer