Imagine trying to run a modern video game on a computer from the 90s; it just won't work. This project updates seven massive scientific programs so they can run on the world's most powerful new supercomputers. It's like upgrading the engine of a car to handle a thousand times more speed without breaking down.
Current scientific simulation software cannot handle the extreme complexity and scale of new exascale supercomputers. This leads to wasted computing power and an inability to process the massive 'data torrents' produced by these machines.
Re-engineered versions of 7 lighthouse astrophysical codes (including Open GADGET and RAMSES) and a set of machine-learning and visualization tools for big data analysis.
If you are a provider dealing with inefficient software on heterogeneous hardware — this project developed re-engineered codes for 7 target applications that optimize energy efficiency and performance. This allows you to offer better compute services for complex physics simulations.
If you are a vendor dealing with outdated code that cannot scale to new hardware — this project developed innovative programming solutions and libraries. This enables you to modernize legacy simulation tools for the exascale era.
If you are a firm dealing with a torrent of simulation data that is too large to process — this project developed machine-learning and visualization tools for high-performance analysis. This reduces the time needed to extract insights from massive datasets.
Based on available project data, no commercial pricing is listed as this is an EU-funded research project with a contribution of EUR 3,997,406.
Yes, the project specifically targets exascale and post-exascale computing capabilities, focusing on scalable parallel and distributed codes.
The project emphasizes the reuse and sharing of algorithms and the adoption of FAIR principles, suggesting a move toward open or standardized software components.
It uses a co-design methodology involving hardware manufacturers and software developers to ensure compatibility with heterogeneous architectural complexity.
The project runs from 2023-01-01 to 2026-12-31.
The consortium is well-balanced for a technical project, consisting of 15 partners across 8 countries. It maintains a 20% industry ratio with 3 industrial partners, including 3 SMEs, which ensures that the 6 universities and 6 research centers are grounded in practical, commercial hardware and software requirements.
Contact Universita Degli Studi di Torino
Contact us to find the specific re-engineered code library that fits your HPC needs.
