Pythonic code vs. Performance ukasz K kol Who am I? Senior - PowerPoint PPT Presentation

Jun 27, 2023 •114 likes •238 views

Pythonic code vs. Performance ukasz K kol Who am I? Senior Python Developer @ STX Next Speaker @ PyCon UK, PySS and local meetups Optimization enthusiast Agenda Introduction Demo! Summary Q&A Introduction

Pythonic code vs. Performance Ł ukasz K ą kol
Who am I? ● Senior Python Developer @ STX Next ● Speaker @ PyCon UK, PySS and local meetups ● Optimization enthusiast
Agenda ● Introduction ● Demo! ● Summary ● Q&A
Introduction
Performance aspects ● CPU time ● Memory ● I/O operations
You can into performance ● Plan ● Measure ● Monitor ● Identify bottlenecks and quick-wins ● Optimize them all! (?)
Profiling tools ● cProfile ● memory_profiler ● sys ● dis
DEMO!
Summary
Summary ● Predict ● Don’t trust presumptions ● Profile everything ● Calculate ROI
Q&A lukasz.kakol@stxnext.pl lukkol.pl/ep2018-performance github.com/lukaszkakol/python-performance goo.gl/gPDTuN

Recommend

Pythonic Coding Style C-START Python PD Workshop C-START Python PD Workshop Pythonic Coding

Pythonic Coding Style C-START Python PD Workshop C-START Python PD Workshop Pythonic Coding Style A Foolish Consistency is the Hobgoblin of Little Minds Guido van Rossum (creator of Python) makes a point: code is read more often than it is

797 views • 20 slides

How to be Pythonic? Design a Query Language in Python Cheuk Ting Ho https://cheuk.dev

Grab the slides: bit.ly/be-pythonic How to be Pythonic? Design a Query Language in Python Cheuk Ting Ho https://cheuk.dev Cheukting @cheukting_ho https://www.twitch.tv/cheukting_ho What does Pythonic mean? (Is it a thing?) What does

576 views • 45 slides

Pythonic Coding Style A Foolish Consistency is the Hobgoblin of Little Minds Guido van Rossum

Pythonic Coding Style A Foolish Consistency is the Hobgoblin of Little Minds Guido van Rossum (creator of Python) makes a point: code is read more often than it is written , so readability counts . Python is one of the few languages with an

555 views • 20 slides

Designing a Pythonic Interface @honzakral Illustrated guide to this import this Disclaimer

Designing a Pythonic Interface @honzakral Illustrated guide to this import this Disclaimer Personal opinions Do as I say, not as I do API API is a service for "code" Fulfills a "contract" Vaguer Simplifying Access

500 views • 34 slides

A PYTHONIC FULL-TEXT SEARCH PAOLO MELCHIORRE ~ @pauloxnet Paolo Melchiorre CTO @ 20tab

A PYTHONIC FULL-TEXT SEARCH PAOLO MELCHIORRE ~ @pauloxnet Paolo Melchiorre CTO @ 20tab Remote worker Software engineer Python developer Django contributor Pythonic >>> import this Beautiful is better than ugly .

891 views • 41 slides

std::map<Code,Performance> myMCU{?} @DanielPenning The mapping between Code &

std::map<Code,Performance> myMCU{?} @DanielPenning The mapping between Code & Performance www.embeff.com World Map (1459) World Map (1525) People admitted they dont know. @DanielPenning The mapping between Code &

470 views • 26 slides

The effect of sampling frequency and front-end bandwidth on the DLL code tracking performance Vi

The effect of sampling frequency and front-end bandwidth on the DLL code tracking performance Vi Vinh T T r ran, Thuan Nguyen, Nagaraj Shivaramaiah, Andrew Dempster Contents GNSS receiver model Local code generator and residual code

500 views • 22 slides

Performance Scaling How is my parallel code performing and scaling? Performance metrics

Performance Scaling How is my parallel code performing and scaling? Performance metrics Measure the execution time T how do we quantify performance improvements? Speed up typically S(N,P) < P Parallel efficiency

377 views • 17 slides

Pseudo An algorithm design language Overview - Concise Pythonic syntax - Rich list support -

Pseudo An algorithm design language Overview - Concise Pythonic syntax - Rich list support - Robust type inference - On-demand objects Compiler Architecture Type Inference Modules L1 Inference Scanner Parser L2 Inference Codegen

591 views • 21 slides

80% of Code Red 2 Code Red 2 re-re- Code Red 1 and Code Red 2 Code Red 2 re- cleaned up

Slammer s Bandwidth-Limited Growth 80% of Code Red 2 Code Red 2 re-re- Code Red 1 and Code Red 2 Code Red 2 re- cleaned up due to released Jan 2004 Nimda endemic dies off released with Oct. onset of Blaster (and 2005; not since

450 views • 34 slides

On Portability, Performance and Scalability of a MPI OpenCL Lattice Boltzmann Code E Calore, S F

On Portability, Performance and Scalability of a MPI OpenCL Lattice Boltzmann Code E Calore, S F Schifano, R Tripiccione Enrico Calore INFN Ferrara, Italy 7 th Workshop on UnConventional High Performance Computing August 26, 2014 Porto,

436 views • 24 slides

Performance metrics How is my parallel code performing and scaling? Performance metrics A

Performance metrics How is my parallel code performing and scaling? Performance metrics A typical program has two categories of components - Inherently sequential sections: cant be run in parallel - Potentially parallel sections ( ) ) = T

408 views • 22 slides

Moving to performance based energy code December 11, 2013 Ian Finlayson Deputy Director, Energy

Creating A Cleaner Energy Future For the Commonwealth Moving to performance based energy code December 11, 2013 Ian Finlayson Deputy Director, Energy Efficiency Division Outline Energy Context in MA Why Performance based code ?

176 views • 17 slides

Evaluation of a performance portable lattice Boltzmann code using OpenCL Simon McIntosh-Smith

Evaluation of a performance portable lattice Boltzmann code using OpenCL Simon McIntosh-Smith Dan Curran Computer Science University of Bristol Twitter: @simonmcs 1 Motivation Our BUDE molecular docking code turned out to show strong

525 views • 17 slides

A Code Curriculum at Your Fingertips: Improving Performance and Resident A Code Curriculum at Your

A Code Curriculum at Your Fingertips: Improving Performance and Resident A Code Curriculum at Your Fingertips: Improving Performance and Resident Confidence in Pediatric Codes Confidence in Pediatric Codes Alyssa Buono, MD-R, Kelli Davis, DO-R,

325 views • 7 slides

The Need for Tuning (1 of 2) You dont need to tune your code! Most important Code

2/21/2012 The Need for Tuning (1 of 2) You dont need to tune your code! Most important Code that works Most important Code that is clear, readable Performance Tuning It will be re-factored It will be modified by

427 views • 7 slides

SWIRL++ :Evaluating Performance Models System Overview to Guide Code Transformation in Model

SWIRL++ T. Rusira et al LCPC19 Convolution SWIRL++ :Evaluating Performance Models System Overview to Guide Code Transformation in Model Guided Optimization Convolutional Neural Networks Code Variants Memory Cost Space Pruning

409 views • 29 slides

CFD Example Aims To familiarise yourself with factors that affect code performance

CFD Example Aims To familiarise yourself with factors that affect code performance compiler implementation and platform compiler optimisation options hyper-threading on ARCHER process placement parallel scaling

231 views • 10 slides

Cracking the Coaching Code Simple Strategies to Drive Sales Team Performance Every Day

Cracking the Coaching Code Simple Strategies to Drive Sales Team Performance Every Day Cracking the Coaching Code Simple Strategies to Drive Sales Team Performance Every Day Becc Holland Mark McWatters Head of Business Development VP of

721 views • 18 slides

Code Analysis Tools and Tips (How to make your code ROCK!) This Presentation is About About

Darryl Parks Code Analysis Tools and Tips (How to make your code ROCK!) This Presentation is About About Code Analysis, not Run-Time monitoring This Presentation is NOT about Performance Analysis T ools Profiling Jconsole or other

913 views • 48 slides

Performance analysis Agenda Code Profiling Linux tools GNU Profiler (Gprof)

Blue Gene/Q User Workshop Performance analysis Agenda Code Profiling Linux tools GNU Profiler (Gprof) bfdprof Hardware Performance counter Monitors IBM Blue Gene/Q performances tools Internal mpitrace Library

1.03k views • 38 slides

Use of SIMD Vector Operations to Accelerate Application Code Performance on Low-Powered ARM and

Use of SIMD Vector Operations to Accelerate Application Code Performance on Low-Powered ARM and Intel Platforms Gaurav Mitra 1 Beau Johnston 1 Alistair P. Rendell 1 Eric McCreath 1 Jun Zhou 2 1 Research School of Computer Science, Australian

572 views • 39 slides

Adult Education Administrative Code Revisions Division of Academics and Performance Erskine

Adult Education Administrative Code Revisions Division of Academics and Performance Erskine Glover, Acting Director Office of Career Readiness Please note that this PowerPoint presentation has been modified from its original version to be

253 views • 12 slides

Motivation for Multithreaded Architectures Processors not executing code at their hardware

Motivation for Multithreaded Architectures Processors not executing code at their hardware potential late 70s: performance lost to memory latency 90s: performance not in line with the increasingly complex parallel hardware as

591 views • 32 slides