Winners Announced for the oneAPI Great Cross-Architecture Challenge
Intel announced the winners of the oneAPI Great Cross-Architecture Challenge in collaboration with the European Organization for Nuclear Research (CERN) and Argonne National Laboratory. The challenge attracted participants from 52 countries across five continents, showing the growing momentum of oneAPI’s cross-architecture, multi-vendor, and open approach. The entrants used oneAPI and Data Parallel C++ (DPC++) to create a variety of applications in domains such as bioinformatics, cryptography, data analytics, education, financial services, genomics, healthcare, image processing, mathematics, molecular dynamics, particle physics, and ray tracing.
The oneAPI Great Cross-Architecture Challenge asked professional and student software developers to use oneAPI to create fast, efficient, and future-ready heterogeneous applications that take full advantage of various XPUs including CPUs, GPUs, FPGAs, and other accelerators. Using the free access to the Intel oneAPI Toolkits and the Intel® DevCloud, which provides the ability to test code and workloads across Intel XPUs, participants had the option of porting an existing C/C++ or CUDA application using the Intel® DPC++ Compatibility Tool or creating an entirely new oneAPI application.
This challenge really showcases the ease of use and freedom of choice that oneAPI’s open, cross-architecture programming model delivers. The participants were able to either quickly port or develop from scratch applications with real-world impact across a range of disciplines. We are highly impressed with the innovative and creative submissions received from around the world, and the positive feedback and growing adoption for oneAPI.
Jeff McVeigh, vice president, Datacenter XPU Products and Solutions at Intel
The participants in the Great Cross-Architecture Challenge demonstrated the potential of oneAPI. Through its use, they were able to write code for heterogeneous hardware architectures with a diverse range of applications. Opening the Intel development environment leveled the playing field for this competition. People from across the world were able to access cutting-edge technology through this developer challenge. We look forward to welcoming the winners of the competition to CERN.
Maria Girone, Chief Technology Officer, CERN openlab
The winning student submissions receiving the internship award include:
- Eugenio Marinelli of France leveraged oneAPI’s complete set of cross-architecture libraries and tools to efficiently develop a new application for implementing scalable, heterogeneous parallel processing algorithms that can be used to quickly and accurately decode digital data stored in synthetic DNA.
- Andrew Pastrello of Australia showed the ease of porting CUDA code to DPC++ by optimizing a music production tool to synthesizing audio from gravitational waveforms produced by binary black hole inspiral-merger-ringdown simulations.
Professional developers receiving the opportunity to take a special tour of CERN include:
- Rafael Campos of Portugal demonstrated oneAPI’s fast and efficient development by adapting OpenCL applications using modern constructs and minimal programming effort. The result is the boosting of performance and power efficiency of bioinformatic applications, specifically for epistasis detection.
- Zhen Ju of China showcased the migration of a CUDA-based application and the benefits of an open programming model for all architectures. The ported application offered a more efficient and accurate solution to filter out redundant sequences in genetic data.
- Ricardo Nobre of Portugal used the Intel DPC++ Compatibility Tool to seamlessly port a CUDA-based application, with more than 95% of their hand-tuned code automatically migrated. The application features collaborative utilization of CPU and GPU devices to find new associations between genotypes and phenotypes.
Participants had access to additional free resources such as code samples, developer guides, webinars, and the DevMesh collaboration portal.
The contest offered more than $40,000 in prizes as well as once-in-a-lifetime opportunities like trips to CERN, a CERN openlab internship, and a chance to work on a project with Argonne National Laboratory. Five grand prize winners were selected by a panel of six esteemed judges:
- Maria Girone, Chief Technology Officer, CERN
- Erik Lindahl, Professor of Physics, Stockholm University
- Simon McIntosh-Smith, Professor of High-Performance Computing, Bristol University
- Heidi Poxon, Distinguished Technologist, Hewlett Packard Enterprise
- Katherine Riley, Director of Science, Argonne National Laboratory
- Michael Wong, Distinguished Engineer, Codeplay Software
Entries were evaluated based on innovation, impact on humanity, use of cross-architecture computing, level of coding expertise, and quality of project explanation. The top five winners were awarded one of the following grand prizes1 :
- One of three trips to CERN for a special tour, or $5,000 cash
- A summer CERN openlab internship (in person or virtual), or $8,000 in cash
- Participate in a oneAPI-related project at Argonne National Laboratory (in person or virtual), or $8,000 in cash
In addition to the grand prizes, 20 contestants received $500 cash prizes for their quality submissions.
Since 2019, oneAPI ecosystem support has steadily grown. More than 60 leading research organizations, companies, and universities support the oneAPI initiative. Their success using Intel oneAPI toolkits is noted in oneAPI ecosystem support and reviews site. A new oneAPI applications catalog details more than 230 applications powered by oneAPI. The following resources are available to help developers build high-performance, cross-architecture applications using oneAPI and the Intel oneAPI Toolkits:
- Key Links: Documentation, oneAPI programming guide, and code samples, and free training.
- Support: For technical support using the Intel oneAPI Toolkits, developers can access the free community forums. Priority Support with direct, private interactions with Intel engineers is included in toolkit commercial packages.
- Training: Free training is available via webinars, deep dive workshops, full learning paths, and more. For customers needing assistance accelerating HPC and AI solutions using oneAPI, a group of specially trained companies offer consulting through a new Intel® oneAPI Technology Partner program.
1 Contingent upon COVID-19 travel restrictions