Відео

Faster ETL Pipeline with Bodo: A Compiler Based Parallel Computing for Big Data Analytics The growing scale and complexity of data and ML workloads demands enormous compute power. Bodo is a new compute platform that aims to improve efficiency and performance of ETL pipelines through automated parallelization of native Python and SQL workloads. Bodo’s resource efficiency and extreme performance ...

In recent years, the number of software services that are running on datacenter and cloud computing platforms has increased rapidly. Datacenters are expensive to operate and represent a significant portion of the world's energy consumption. Efforts to improve the efficiency of software that runs on datacenters can greatly reduce the energy use and costs of operations and improve the responsiven...

Modern hardware is highly parallel, but not only in terms of multiprocessing. There are many other forms of parallelism that, if used correctly, can greatly boost program efficiency - and without requiring more CPU cores. One such type of parallelism actively adopted by CPUs is "Single Instruction, Multiple Data" (SIMD): a class of instructions that can perform the same operation on a block of ...

ByteDance in-house Redex passes A handful Redex passes have been developed in ByteDance in addition to existing Redex passes. We'd like to share some of them, RenameClass, StringBuilderOutliner, KtDataClass. These three passes saved about ~1M, ??%. *Exploring ART Optimization Complement * ART optimization is triggered during both jit and dex2aot, however both are running with restricted computa...

Most modern performance work is centered around technologies and tools. Following Jeff Bezos' advice, this talk, instead of trying to predict the future, covers some of the oldest mathematical principles and ways they've shaped modern software. Parallel and distributed computing, caching and proxy computations are just a few examples of leveraging associativity, idempotence, Kleisli category an...

Personal computers and devices are unbelievably fast but often struggle to keep up with mundane tasks. Why hasn't software managed to scale in performance with hardware? This talk looks at some of the reasons, the characteristics of fast systems, the limits of human perception, and the need to rethink how software is authored to build efficient systems. skillsmatter.com/skillscasts/17836-where-...

When it comes to software performance, ideally we want two things: 1. The performance of our program doesn’t change as the features are added. 2. The runtime of the program grows linearly to the data set size. But this often fails to happen. As new features are added or the dataset grows, the software starts running much slower than one would expect. In this talk we will explore what are the fa...

Join Luca Chiabrera, Paul McKenney, Daniele Salvatore Albano, Moritz Beller, and our host lonut ZoIti for a thought provoking conversation of what's next in the Computing Performance world !

Dataproc is a fully managed and highly scalable Google service that facilitates quickly deploying clusters and executing Spark applications. Nevertheless, the user remains responsible for sizing the Dataproc infrastructure and defining the Spark application execution parameters. These activities have a relevant impact on both the execution time of the Spark applications and the cost of the Data...

Although there is no magic wand that can wish away all bugs in concurrent software, there is a wealth of tools, techniques, and experience that permit moving towards that goal. This talk takes on the viewpoints of semantics, software engineering, installed base, software stack, and finally natural selection in order to name the price of fast and robust concurrent software. Performance Summit Se...

cachegrand is a super-scalar caching platform, aiming initially to be a Redis drop-in replacement, to achieve high performances a lot of components have been written from the scratch with performance in mind, among other things the internal parallel hashtable, built only to store 64 bit numbers, is using a number of different techniques to maximize the throughput: data oriented design to take a...

Imagine what we could do if we could predict performance regressions before they actually impact production, at code authorship time? In this talk, I will set the stage of the performance prediction landscape, will clarify why predicting performance regressions of a piece of code is difficult, go over the many things we learned as part of the SuperPerforator initiative at Facebook, and share ou...

Join Stefano Doni, Ivica Bogosavljevic, Yaniv Sabo, Simone Casagranda, Raman Pandey, Nian Sun and our host lonut ZoIti for a thought provoking conversation of what's next in the Computing Performance world !

Connected vehicle platform is one of the important performance sensitive platforms, where the role of performance engineer is to assess the IoT devices which are connected to the cloud solution and communicate via telemetries. Performance Summit September 2021

Using ML to Automatically Optimize Kubernetes for Cost Efficiency and Reliability - Stefano Doni

Performance Optimization Techniques for Mobile Devices: An Overview - Ivica Bogosavljevic

Task scheduling optimizations in Mobile Application - Nian Sun

Shipping 2021 user experiences on 2011 phones - Simone Casagranda & Yaniv Sabo

Performance Summit Trailer Sep 2021

London Perf Summit - Panel Discussion

We Should Become Good at Optimizing our Code - Denis Bakhvalov

Understanding and Measuring CPU throttling in containerized environments - Francesco Fabbrizio

How AI optimization will debunk 4 long-standing Java tuning myths - Stefano Doni

Virtual machine warmup blows hot and cold - Laurence Tratt

Debugging performance issues with Java Flight Recorder - Alexander Kachur

On access control model of Linux native performance monitoring - Alexei Budankov

Query-based compiler architectures - Olle Fredriksson

nanoBench: A Low-Overhead Tool for Running Microbenchmarks on x86 Systems - Andreas Abel

Automatically avoid the top performance patterns in distributed systems - Andreas Grabner