Sharing is caring. Join us on Slack.
This talk will review two especially fun/interesting incidents. One is from a pre-production situation and the other a live SEV. In both cases unexpected and interesting data analysis led to a great result. These kinds of Sherlock Holmes moments are some of the most fun we get to have doing Software Performance Engineering.
Rico Mariani | Software Engineer, Facebook
This talk presents ugrep, a high-performance grep tool. The internals of ugrep will be revealed, including a new multi-string pattern match algorithm, fuzzy regex pattern matching, task-parallel compressed tar/zip archive search, and optimal load balancing with lock-free job queues.
Dr. Robert van Engelen | Founder and CEO, Genivia Inc
Data structures are everywhere. They define the behavior of modern data systems and data-driven algorithms. For example, with data systems that utilize the correct data structure design for the problem at hand we can reduce the monthly bill of large-scale NoSQL applications on the cloud by thousands of dollars. We can accelerate data science tasks by being able to dramatically speed up the computation of statistics over large amounts of data. We can train drastically more neural networks within a given time budget, improving accuracy.
However, knowing the right data structure and system design for any given scenario is a notoriously hard problem; there is a massive space of possible designs while there is no single design that is perfect across all data, queries, and hardware scenarios. We will discuss our quest for the first principles of data structures and system design. We will show signs that it is possible to reason about the design space, heading toward a future of self-designing data systems.
Stratos Idreos | Professor, Harvard University
With an exponential growth in the number of features being shipped to production, how as a performance engineering team can we empower developers to be more closer to performance of the feature being shipped?
We will discuss some of the process and technology which is being used by our Performance Engineering Team to help us shift more leftwards.
Vivek Koul | Engineering Leadership, Mcgraw-Hill
The mantra for perf is measure, measure, and measure. As perf owners, do we apply that ourselves? Do we measure how successful our programs are? Most of us measure the number of egregious perf incidents that hit our customers - but that's just the headline number. What are the numbers we need to capture to find out if we (perf owners) are effective? That's this talk.
Mani Ramaswamy | Technical Program Manager, Facebook
As systems get more complex, reasoning about performance gets more difficult. Telemetry data emitted by our services is noisy and usually unhelpful in stressful situations. Distributed Tracing, in particular, can provide rich, contextual data but root-cause analysis can still be convoluted. In this talk, I'll review a few statistics-based approaches we have applied to help quickly identify which properties of the system are correlated with performance issues.
In order to support this type of aggregate trace analysis, we need data, but data isn't cheap. We want to gather only the relevant traces and bias towards traces that have abnormal behavior. I'll also talk about a few sampling approaches we use for analysis to minimize cost and overhead.
Kalyana Chadalavada | Software Engineering Manager, Google
At perf.dev, we are building a platform to give companies complete confidence in their mobile performance strategy. Usability and sophistication of developer debugging tools are critical to the accurate diagnosis of performance problems. Our debugging emulator provides the most precise method tracing in the industry. We’ll dive into how the tool works and give a short demo of the emulator in action.
Leland Takamine | Cofounder & CEO, perf.dev
In this talk, I will describe the performance improvements work that .NET Runtime team has done in .NET 5 for ARM64 target. I will cover our investigation strategy and methodology in achieving high performance gains. The talk will cover most of the things we have recently published in our devblog.
Kunal Pathak | Software Engineer, Microsoft
Redpanda: A new storage engine built from the ground up for Kafka-API compatibility with Kernel bypass technologies for 10x lower tail latencies
In this talk we’ll cover how we built a new storage engine from scratch using a thread per core architecture for predictable tail latencies. We’ll dive deep into the challenges of adapting a well known protocol (Kafka-API) on top of this new storage engine and the performance gains of optimizing for modern hardware.
Alexander Gallego | Founder, vectorized.io