2022-11-08 Intel OneAPI Training (hoap1w22)

CourseIntel OneAPI Training
Numberhoap1w22
Places available12
Date08.11.2022 – 10.11.2022
Price€ 0.00
PlaceONLINE
Room
Registration deadline07.11.2022 12:00
E-maileducation@lrz.de

Contents

The oneAPI initiative (www.oneapi.com) is a cross-industry, open, standards-based unified programming model that delivers a common developer experience across multiple / multi-vendor CPU and  accelerator architectures.
It is designed for faster application performance, more productivity, and greater innovation. The oneAPI industry initiative encourages collaboration on the oneAPI specification and compatible oneAPI implementations across the ecosystem.

Intel released the Intel® oneAPI Toolkits implementing its own programming languages, models, libraries and tools that are built to the above mentioned oneAPI specifications targeting Intel CPUs and accelerators (GPU and FPGAs).

The Intel® oneAPI Base & HPC Toolkit solution – the successor to the Intel® Parallel Studio XE tools suites – provides

  1. high performance LLVM based compilers for C/C++, Fortran,
  2. OpenMP for offloading purposes,
  3. performance optimised libraries such as Intel® oneAPI Math Kernel Library and Intel® MPI Library and
  4. analysis tools (VTune Profiler and Advisor) enhanced to support heterogeneous development.

We welcome you to join us in learning about the oneAPI initiative and how to get on a standards-based path for heterogeneous programming.
This 3 half-days online workshop will teach the participants about the new and expanded features of the Intel oneAPI Toolkits.
The presenters will show you how to use the tools for shared and distributed computing on heterogenous hardware platforms including Intel CPUs, Intel hardware accelerators and Intel discrete graphics solutions.

After the event the attendees should understand:

  • the transition from the Intel Parallel Studio Development tool to Intel oneAPI Toolkits,
  • how to use the Intel oneAPI Toolkits to develop heterogenous applications running on CPUs and HW accelerators,
  • make use of the offered development environments such as the Intel Dev Cloud,
  • migrate non oneAPI heterogenous code (CUDA) to oneAPI programming models (SYCL),
  • how to start developing and optimising the performance with the oneAPI development environment

The workshop is co-organised by LRZ and Intel for EuroCC@GCS, the German National Competence Centre for High-Performance Computing.


Preliminary Agenda

All times are in CET.

Day 1: Tuesday


 Start

 End

 Duration

Topic

 10:00

 10:15 

 00:15

Welcome and introduction to Day 1

 10:15

 10:35

 00:20

OneAPI – Introduction to a new development environment
- Concept and oneAPI standardisation initiative
- Intel’s Tools Implementation – Intel oneAPI Toolkits and libs
- Transition from Intel Parallel Studio XE to Intel oneAPI Toolkits

 10:35

 10:55

00:20

Introduction to the DevCloud
- Purpose: demoing, testing and porting applications
- Hardware and software offerings
- How to onboard & how to get an DevCloud account

 10:55

 11:55

 01:00

Direct programming with oneAPI Compilers (Part 1) – with live demos

- Intro to heterogenous programming model with SYCL 2020
- SYCL features and examples
   o  “Hello World” example
   o  Device selection
   o  Execution model

 11:55

 12:55

 01:00

Lunch

 12:55

 13:10

 00:15

Using oneAPI on SuperMuc-NG

 13:10

 14:40

 01:30

Direct programming with oneAPI Compilers (Part 2) – with live demos

   o  Compilation and execution flow 
   o  Memory model; buffers, Unified Shared Memory (USM)
   o  Performance optimisations with SYCL features

 14:40

 14:45

 00:05

Wrap up





Day 2: Wednesday

 Start

End

Duration

Topic

 10:00

 11:00

 01:00

Intel OpenMP for Offloading – with live demos
- Parallelising heterogenous applications with OpenMP 5.1
- Mixing of OpenMP and SYCL

 11:00

 11:35

 00:35

Intel oneAPI libraries (oneMKL) for HPC  – with live demos
- Performance optimised libraries for numerical simulations and other purposes

 11:35

 12:05

 00:30

Intel Debugging Tools for heterogenous programming (CPU, GPU ) – with live demos

 12:05

 13:05

 01:00

Lunch

 13:05

 13:45

 00:40

Open Source Compatibility tool for porting purposes (SYCLomatic) – with live demos
- Migration CUDA based GPU Applications to SYCL

 13:45

 14:25

 00:40

Dynamic debugging with Intel Inspector - with live demos
- Identifying memory and threading errors (data races and deadlocks)

 14:25

 14:55

 00:30

Q&A





Day 3: Thursday

 Start

End 

Duration

Topic

 10:00

 11:00

 01:00

A 3rd party oneAPI case study: Gromacs – a molecular dynamics engine
- Heterogenous design consideration, alternatives and comparisons
- Real scheduling
- SYCL – oneAPI and other implementations
- SYCL in Gromacs 2022

 11:00

 12:15

 01:15

Application profiling for heterogenous hardware – with live demos
- Profiling tools interfaces for GPU – open source lightweight tools
- Profile heterogenous SYCL/OpenMP workloads with Intel VTune profiler
- Share experiences/key findings with Gromacs related porting and optimisation efforts

 12:15

 13:15

 01:00

Lunch

 13:15

 14:30

 01:15

Application profiling for heterogenous hardware – with live demos
- Estimate performance potential gains with Offload Advisor ( CPU -> HW accelerator)
- Analyse heterogenous SYCL/OpenMP Workloads with Intel Advisor and roofline analysis

 14:30

 14:55

 00:25

Programming for distributed HPC systems using Intel MPI

 14:55

 15:25

 00:30

Q&A  –  Wrap up





Prerequisites

It is assumed that participants have a basic knowledge about HPC application/software programming and should be able to program in C/C++ or Fortran.

Hands-On

There will be only live demos.

Language

English

Lecturers

Heinrich Bockhorst, Gennady Fedorov, Rafael Lago, Klaus-Dieter Oertel, Edmund Preiss, Alina Shadrina, Dmitry Sivkov, Igor Vorobstov (Intel), Andrey Alekseenko (KTH), Nisarg Patel (LRZ)

Prices and Eligibility

The course is open and free of charge for people from academia and industry from the Member States (MS) of the European Union (EU) and Associated Countries to the Horizon 2020 programme.

Registration

Please register with your official e-mail address to prove your affiliation.

Withdrawal Policy

See Withdrawal

Legal Notices

For registration for LRZ courses and workshops we use the service edoobox from Etzensperger Informatik AG (www.edoobox.com). Etzensperger Informatik AG acts as processor and we have concluded a Data Processing Agreement with them.

See Legal Notices


No.

Date

Time

Leader

Location

Room

Description

108.11.202210:00 – 14:45
ONLINE
Intel OneAPI Training
209.11.202210:00 – 14:00
ONLINE
Intel OneAPI Training
310.11.202210:00 – 15:30
ONLINE
Intel OneAPI Training