Join us for this year’s oneAPI Developer Summit at IWOCL focused on oneAPI and SYCL for accelerated computing across xPU architectures (CPU, GPU, FPGA, and other accelerators). In this two-day virtual conference, you will hear from industry and academia speakers working on innovative cross-platform, multi-vendor architecture solutions developed on oneAPI. Learn from fellow developers and connect with other innovators. Please join us, a self-sustained, vibrant community to support each other using oneAPI, SYCL and Data Parallel C++.
9:00 - 9:15 CT
9:15 - 9:45 CT
Parallel Programming Models and SYCL in the UK’s Exascale Programs
The UK is investing heavily in getting its science codes ready for the Exascale era. The ExCALIBUR project is a 5-year, £45M program focused on codes and algorithms, getting them ready for Exascale architectures. A key issue for ExCALIBUR is what parallel programming models to adopt, and so in this talk, we will examine these drivers, and explore how parallel C++ models such as SYCL are being evaluated and adopted.
9:45 - 10:15 CT
Learning SYCL: Challenges and Opportunities
The EuroHPC Joint Undertaking (EuroHPC-JU) aims at developing exascale computing capabilities within the EU and promoting their use, both in academia and in industry. In order to harness such resources efficiently, a network of national competence centers is established to foster awareness for HPC and develop advanced HPC competence. Given that increasing portions of computational power are delivered through accelerators, it is critical to develop and spread knowledge of performant and performance-portable ways of building next-generation HPC applications. As the EuroCC node for Sweden, we are pioneering learning resources for SYCL. This contribution will showcase our pedagogical approach: how it is well-suited for remote teaching and self-learning. I will discuss and motivate the structure of our SYCL lesson, how it is delivered in practice, and how it can be freely reused.
10:15 - 10:30 CT Break
10:30 - 10:45 CT
TAU Performance System for SYCL and E4S
The TAU Performance System® is a versatile profiling and tracing toolkit included in the Extreme-Scale Scientific Software Stack(E4S). TAU supports performance evaluation of HPC and AI/ML applications built running on Intel CPUs and GPUs using Intel oneAPI. TAU supports performance instrumentation of un-modified binaries as well as compiler-based instrumentation. E4S provides a curated, Spack package manager based distribution of open-source libraries, tools (including TAU), and data and visualization packages available for installation on bare-metal or Singularity or Docker containers.
11:00 - 11:30 CT
Challenges Faced When Porting Ginkgo to the SYCL Ecosystem
We present our experience of a platform-portable sparse linear algebra ibrary – Gingko. Previously, we already ported essential parts of our library like SpMV or Krylov solvers to oneAPI based on sycl-1.2.1. We present challenges and our solutions when porting functionality and in stayin up to date due to the changed SYCL 2020, issues with per-kernel code splitting, and supporting varying hardware. We generate kernels suitable for idfferent devices and give our approach of pre-compiled kernels. To wrap up, we show some experimental results of preconditioners or Algebraic Multigrad.
11:30 - 11:45 CT Break
11:45 - 12:30 CT
Developing for the Next Generation of RISC-V Processors with SYCL
The growth in adoption of RISC-V by an increasing number of companies is coinciding with a demand from software developers for processors that can help them deploy cutting edge AI and HPC applications. How do we enable software developers to take advantage of the latest RISC-V processors, in particular those designed with AI processing in mind? One crucial piece is a programming model that can be used to achieve portability across RISC-V processors, and SYCL is quickly bringing a solution for this.
This panel session will explore the latest developments from companies involved in bringing the latest RISC-V processors to market alongside exploring the needs of researchers and developers writing the next generation of AI and HPC software.
Mark Himelstein – Download Presentation Deck
Zdenek Prikryl – Download Presentation Deck
Hideki Sugimoto – Download Presentation Deck
Michael Wong – Download Presentation Deck
12:30 - 12:45 CT
12:45 - 1:45 CT
9:00 - 9:15 CT
9:45 - 10:45 CT
oneAPI SYCL Essentials
• Overview of oneAPI
• Introduction to the Intel® DevCloud
• Introduction to Jupyter notebooks used for training
• Overview of DPC++
• Introduction to DPC++
• Vector-add coding exercise
• Intel® VTune™ Profiler and Intel® Advisor exercise
The participant will come away with an understanding of:
• How oneAPI can solve the challenges of programming in a heterogeneous world
• Use oneAPI solutions to enable your workflows
• Experiment with oneAPI tools and libraries on the Intel® DevCloud
• Understand the Data Parallel C++ (DPC++) language and programming model