Thematic CERN School of Computing 2023

Jun 11, 2023, 11:00 AM → Jun 17, 2023, 11:30 AM Europe/Zagreb

MedILS, Split, Croatia

Alberto Pace (CERN), Kristina Gunne (CERN), Jarek Polok (CERN)

The 12th Thematic CERN School of Computing (tCSC 2023) will take place on June 11-17, 2023.The theme of the school is "Scientific Software for Heterogeneous Architectures" - see the academic programme for more details.

This school is organized by CERN in collaboration with the University of Split. The school will take place in Split, Croatia, and be hosted at the Mediterranean Institute For Life Sciences (MEDILS) Conference Centre. The Centre is a historical renovated building situated in a wooded and landscaped park located on the Adriatic Sea coast, a few kilometers from the centre of Split. 

The School is aimed at postgraduate (i.e. minimum of Bachelor degree or equivalent) students, engineers and scientists with a few years' experience in particle physics, in computing, or in related fields. We welcome applications from all countries and nationalities.

Applications are closed.

Important Dates

• Monday March 13- applications deadline
• Tuesday April 4 - invitations sent to the selected participants
• Tuesday 2 May - registration fee payment deadline
• Sunday June 11 to Saturday June 17 - the school

              

Survey test

tCSC 2023 Booklet

tCSC 2023 Logistics information

tCSC 2023 Poster

Sun, June 11

12:20 PM → 1:30 PM Lunch

1:30 PM → 4:00 PM Registration

4:00 PM → 4:20 PM Welcome and introduction

Welcome-to-tCSC-2023.pdf

4:20 PM → 5:00 PM Self presentations

One minute-one slide self presentations

Self-presentations-tCSC-2023.pdf

5:15 PM → 5:30 PM Transport to Split

5:30 PM → 7:30 PM Visit of Split old town

7:30 PM → 9:30 PM Welcome Dinner - Konoba Varos

The dinner will take place at the Konoba (Tavern) Varos
Tavern "Varos"
Ban Mladenova 9
21 000 Split
Croatia
Website: http://www.konobavaros.com/hr/

Mon, June 12

7:30 AM → 8:30 AM Breakfast

8:45 AM → 9:45 AM Opening Session

Speakers: Alberto Pace (CERN), Ivica Puljak (University of Split. Fac.of Elect. Eng., Mech. Eng. and Nav.Architect. (HR)), Ivica Puljak (Technical University of Split FESB), Toni Sculac (University of Split Faculty of Science (HR))

2023-06-11-School-Presentation-Generic-Split-v5.pdf

2023-06-11-School-Presentation-Generic-Split-v5.pptx

tCSC 2023 Zoom room

tCSC_about_Split_Puljak_2023.pdf

tCSC_about_Split_Puljak_2023.ppt

9:45 AM → 10:15 AM Coffee break

10:15 AM → 11:15 AM Preparing for the HL-LHC computational challenge

In this talk we will introduce some basic concepts related to HEP data processing and analysis workflows, seeing them in action in the context of LHC experiments. We’ll also talk about the evolution of the LHC accelerator and experiments. We’ll characterise at a high level what are the consequences of those upgrades for the HEP data processing software, in particular in the context of an evolving hardware and computing infrastructure.

Speaker: Danilo Piparo (CERN)

Preparing for the HL-LHC computational challenge

tCSC 2023 Zoom room

11:15 AM → 12:15 PM Introduction to efficient computing

Technologies and Platforms - lecture 1

• The evolution of computing hardware and what it means in practice
• The seven dimensions of performance
• Controlling and benchmarking your computer and software
• Software that scales with the hardware
• Advanced performance tuning in hardware

Speaker: Andrzej Nowak

Introduction to Efficient Computing

12:15 PM → 1:00 PM Lunch

1:00 PM → 2:00 PM Study time or sports

2:15 PM → 3:15 PM Hardware evolution and heterogeneity

Technologies and Platforms - lecture 2

• Accelerators, co-processors, heterogeneity
• Memory architectures, hardware caching and NUMA
• Compute devices: CPU, GPU, FPGA, ASIC etc.
• The role of compilers

Speaker: Andrzej Nowak

Hardware Evolution and Heterogeneity

3:15 PM → 3:45 PM Coffee break

3:45 PM → 5:45 PM Technologies and Platforms - exercise

Speaker: Andrzej Nowak

5:45 PM → 6:45 PM Data-oriented design

Technologies and Platforms - lecture 3

• Hardware vectorization in detail – theory vs. practice
• Software design for vectorization and smooth data flow
• How can compilers and other tools help?

Speaker: Andrzej Nowak

Data Oriented Design

7:15 PM → 8:15 PM Dinner at MEDILS

Tue, June 13

7:30 AM → 8:30 AM Breakfast

8:45 AM → 9:45 AM Writing parallel software

Parallel and Optimised Scientific Software - lecture 1

• Amdahl's and Gustafson's laws
• Asynchronous execution
• Finding concurrency, task vs. data parallelism
• Using threading in C++ and Python, comparison with multi-process
• Resource protection and thread safety
• Locks, thread local storage, atomic operations

Speaker: Sebastien Ponce (CERN)

Writing parallel software

9:45 AM → 10:00 AM Announcements

10:00 AM → 10:15 AM Coffee break

10:15 AM → 11:15 AM Modern programming languages for HEP

Parallel and Optimised Scientific Software - lecture 2

• Why Python and C++ ?
• Recent evolutions: C++ 11/14/17
• Modern features of C++ related to performance
• Templating versus inheritance, pros and cons of virtual inheritance
• Python 3, and switching from Python 2

Speaker: Sebastien Ponce (CERN)

Modern programming languages for HEP

11:15 AM → 12:15 PM Technologies and Platforms - exercise

Speaker: Andrzej Nowak

12:15 PM → 1:00 PM Lunch

1:00 PM → 2:00 PM Study time or sports

2:00 PM → 2:20 PM Technologies and Platforms - exercise debriefing

Speaker: Andrzej Nowak

2:20 PM → 3:20 PM Optimizing existing large codebase

Parallel and Optimised Scientific Software - lecture 3

• Measuring performance, tools and key indicators
• Improving memory handling
• The nightmare of thread safety
• Code modernization and low level optimizations
• Data structures for efficient computation in modern C++

Speaker: Sebastien Ponce (CERN)

Optimizing existing large codebase

3:20 PM → 3:45 PM Coffee break

3:45 PM → 6:45 PM Optimizing existing large codebase - exercise

Optimisation of an existing, production grade large codebase

Speaker: Sebastien Ponce (CERN)

Optimizing existing large codebase - exercise 4

7:15 PM → 8:15 PM Dinner at MEDILS

Wed, June 14

7:30 AM → 8:30 AM Breakfast

8:45 AM → 9:45 AM Scientific computing on heterogeneous architectures

Programming for Heterogeneous Architectures - lecture 1

• Introduction to heterogeneous architectures and the performance challenge
• From general to specialized: Hardware accelerators and applications
• Type of workloads ideal for different accelerators
• Trade-offs between multi-core and many-core architectures
• Implications of heterogeneous hardware on the design and architecture of scientific software
• Embarrassingly parallel scientific applications in HPC and CERN

Speaker: Dorothea Vom Bruch (Aix Marseille Univ, CNRS/IN2P3, CPPM, Marseille, France)

Scientific computing on heterogeneous architectures

9:45 AM → 10:00 AM Announcements

10:00 AM → 10:15 AM Coffee break

10:15 AM → 11:15 AM Practical vectorization

Parallel and Optimised Scientific Software - lecture 4

• Measuring vectorization level
• What to expect from vectorization
• Preparing code for vectorization
• Vectorizing techniques in C++: intrinsics, libraries, autovectorization

Speaker: Sebastien Ponce (CERN)

Practical vectorization

11:15 AM → 12:15 PM Programming for GPUs

Programming for Heterogeneous Architectures - lecture 2

• From SIMD to SPMD, a programming model transition
• Thread and memory organization
• Basic building blocks of a GPU program
• Control flow, synchronization, atomics

Speaker: Dorothea Vom Bruch (Aix Marseille Univ, CNRS/IN2P3, CPPM, Marseille, France)

Programming for GPUs

12:15 PM → 12:20 PM School photo

12:30 PM → 1:15 PM Lunch

1:20 PM → 6:05 PM Excursion to Omis / Cetina - BUS LEAVES AT 13:20

7:00 PM → 8:30 PM Excursion dinner

Thu, June 15

7:30 AM → 8:30 AM Breakfast

8:45 AM → 9:45 AM Performant programming for GPUs

Programming for Heterogeneous Architectures - lecture 3

• Data locality, coalesced memory accesses, tiled data processing
• GPU streams, pipelined memory transfers
• Under the hood: branchless, warps, masked execution
• Debugging and profiling a GPU application

Speaker: Daniel Campora (University of Maastricht)

Performant programming for GPUs

9:45 AM → 10:00 AM Announcements

10:00 AM → 10:15 AM Coffee break

10:15 AM → 11:15 AM Practical vectorization - exercise

Speaker: Sebastien Ponce (CERN)

Practical vectorization - exercise 5

11:15 AM → 12:15 PM Design patterns and best practices

Programming for Heterogeneous Architectures - lecture 4

• Good practices: single precision, floating point rounding, avoid register spilling, prefer single source
• Other standards: SYCL, HIP, OpenCL
• Middleware libraries and cross-architecture compatibility
• Reusable parallel design patterns with real-life applications

Speaker: Daniel Campora (University of Maastricht)

Design patterns and best practices

12:15 PM → 1:00 PM Lunch

1:00 PM → 3:00 PM Study time or sports

3:00 PM → 3:20 PM Optimisation + vectorization - exercise debriefing

Speaker: Sebastien Ponce (CERN)

3:20 PM → 4:40 PM Programming for heterogeneous architectures - exercise

Speakers: Daniel Campora (Nikhef National institute for subatomic physics (NL)), Dorothea Vom Bruch (Aix Marseille Univ, CNRS/IN2P3, CPPM, Marseille, France)

Exercise instructions

4:40 PM → 5:10 PM Coffee break

5:10 PM → 6:30 PM Programming for heterogeneous architectures - exercise

Speakers: Daniel Campora (Nikhef National institute for subatomic physics (NL)), Dorothea Vom Bruch (Aix Marseille Univ, CNRS/IN2P3, CPPM, Marseille, France)

7:15 PM → 8:15 PM Dinner at MEDILS

IESC

Fri, June 16

7:30 AM → 8:30 AM Breakfast

8:45 AM → 10:00 AM Programming for heterogeneous architectures - exercise

Speakers: Daniel Campora (University of Maastricht), Dorothea Vom Bruch (Aix Marseille Univ, CNRS/IN2P3, CPPM, Marseille, France)

10:00 AM → 10:15 AM Coffee break

10:15 AM → 11:30 AM Programming for heterogeneous architectures - exercise

Speakers: Daniel Campora (University of Maastricht), Dorothea Vom Bruch (Aix Marseille Univ, CNRS/IN2P3, CPPM, Marseille, France)

11:30 AM → 11:45 AM Announcements

11:45 AM → 12:15 PM Programming for heterogeneous architectures - exercise debriefing

Speakers: Daniel Hugo Campora Perez (Nikhef National institute for subatomic physics (NL)), Dorothea Vom Bruch (Aix Marseille Univ, CNRS/IN2P3, CPPM, Marseille, France)

12:25 PM → 12:30 PM Thank you MedILS

12:30 PM → 1:15 PM Lunch

1:30 PM → 2:15 PM Exam

Exam logistics

Exam results

2:15 PM → 3:15 PM Summary and future technologies overview

Technologies and Platforms - lecture 4

• Teaching program summary and wrap-up
• Next-generation memory technologies and interconnect
• Future computing evolution

Speaker: Andrzej Nowak

Future Directions and Wrap-up

3:10 PM → 3:55 PM Student Lightning talks

3:15 PM My current research in the field of Big Data

My talk would be about our current research in the field of big data, where we try to implement a system that would scale its speedup to over a hundred nodes. Currently, we are working with astronomers and took the simulation and analysis of pulsars as a first use case for our implementation. My colleague, as a Phd in Physics, came up with the calculations of the simulations in python and I translated it into C++ to increase the performance. This first step is done now and we achieved a runtime increase of over factor one hundred by switching to a lower level programming language. The next task will be to decide on an architecture for the parallel processing and I hoped that presenting this at the tCSC could spark some interesting discussions.

Speaker: Marcel Trattner (HTW Berlin / PUNCH4NFID)

Lightning_Talk_Marcel.pdf

3:25 PM Track reconstruction with Quantum Computing

Quantum Computing is a very promising field that can potentially have a great impact of High Energy Physics. I will give an overview of my PhD project about charged particles track reconstruction using a quantum algorithm.

Speaker: Davide Nicotra (Nikhef National institute for subatomic physics (NL))

QuantumTrackReco.pdf

QuantumTrackReco.pptx

3:35 PM Accelerating the ATLAS Calorimeter Trigger with GPUs

In preparation for the High-Luminosity LHC, the ATLAS Collaboration is interested in evaluating the feasibility of employing hardware acceleration to its software-based high-level trigger. The 16th most costly algorithm in the ATLAS trigger, and the most computationally expensive part of calorimeter reconstruction, is topological clustering, which allows reconstructing the showers that take place in the calorimeters. I would like to present my ongoing efforts to implement a GPU-accelerated alternative to this algorithm, dubbed topo-automaton clustering, and report our current results, with a speed-up of ~6 for di-jet events in run 4 conditions and ~12 for ttbar events in run 3 conditions.

Speaker: Nuno Dos Santos Fernandes (LIP - Laboratorio de Instrumentação e Física Experimental de Partículas (PT))

Accelerating the ATLAS Calo Trigger with GPUs - Nuno Fernandes.pptx

3:55 PM → 4:30 PM Coffee break

4:30 PM → 6:00 PM Free time

6:00 PM → 7:15 PM Closing Session

7:30 PM → 9:30 PM Closing dinner

Sat, June 17

7:30 AM → 8:30 AM Breakfast

9:00 AM → 11:30 AM Departures