OPTIMA · Project

FPGA-Accelerated Computing That Cuts Industrial Simulation Energy Costs by 10x

digitalPilotedTRL 7

Imagine your factory runs heavy computer simulations — fluid dynamics, robotics, structural analysis — and each run eats massive electricity and takes hours. OPTIMA took those same simulations and rewired them to run on a different kind of chip called an FPGA, which is like a Swiss Army knife processor you can reshape for each job. The result? The same calculations finish using a fraction of the energy compared to standard processors. Six SMEs and four research groups across Europe built open-source tools so any company can tap into this, not just tech giants.

By the numbers

10x

Better energy-delay product vs CPU-based systems

Better energy-delay product vs GPU-based systems

Application domains covered

SME partners driving industrial adoption

Consortium partners across 6 countries

The business problem

What needed solving

Companies running heavy industrial simulations — CFD, robotics, structural analysis — face ballooning energy costs and long computation times. Standard CPU clusters are power-hungry, and even GPU acceleration hits efficiency ceilings. As simulation workloads grow, the electricity bill and hardware costs become a serious competitive disadvantage, especially for SMEs that can't afford unlimited cloud compute budgets.

The solution

What was built

OPTIMA delivered final, optimized versions of four industrial applications — SAILFISH-CFD (fluid dynamics), MESHFREE (meshless simulation), finite element method software, and a robotics simulation tool — all running on FPGA-based HPC systems. They also produced open-source libraries and optimized programming environments to help others port their own applications to FPGA hardware. The robotics simulation and libraries are available on GitHub under Apache 2.0.

Audience

Who needs this

Engineering firms running computational fluid dynamics simulationsRobotics companies needing energy-efficient physics simulationStructural engineering consultancies using finite element analysisHPC service providers looking to differentiate with FPGA offeringsManufacturing companies with heavy simulation workloads for product design

Business applications

Who can put this to work

Automotive & Aerospace Engineering

mid-size

Target: Engineering firms running computational fluid dynamics (CFD) simulations

If you are an engineering company spending heavily on CFD simulations for aerodynamics or thermal management — OPTIMA developed SAILFISH-CFD optimized for FPGA-based HPC systems, delivering more than 10x better energy efficiency than CPU-based runs and more than 3x better than GPU-based systems. This means your simulation farm could run the same workloads at a fraction of the electricity cost, or run significantly more simulations within the same power budget.

Robotics & Automation

SME

Target: Robotics companies running physics simulations for motion planning and digital twins

If you are a robotics company that relies on simulation for testing robot behavior before deployment — OPTIMA built a robotics simulation application released as open source under Apache 2.0 on GitHub. This lets you run robotics simulations on energy-efficient FPGA hardware instead of burning through expensive GPU clusters, with more than 10x energy-efficiency improvement over CPUs.

Civil & Structural Engineering

any

Target: Engineering consultancies running finite element analysis for structural design

If you are a structural engineering firm running finite element method (FEM) calculations for building design or crash testing — OPTIMA delivered a final version of FEM software optimized for FPGA hardware, plus the MESHFREE application for meshless simulations. These tools let you run complex structural analyses with dramatically lower energy consumption, cutting your HPC operational costs.

Frequently asked

Quick answers

What would this cost us to adopt?

OPTIMA's key outputs — robotics simulation software, open-source libraries, and programming environments — are released under Apache 2.0 open-source licenses. The software itself is free. Your main costs would be acquiring or renting access to FPGA-populated HPC systems and staff training on the new programming environments.

Does this work at industrial scale or is it still lab-grade?

The project delivered final versions (not just prototypes) of multiple industrial applications: SAILFISH-CFD, MESHFREE, FEM software, and robotics simulation. These were optimized and tested on two FPGA-populated HPC systems across 3 different application domains. With 6 SMEs driving the project, the tools were built for real industrial workloads.

What about IP and licensing?

The robotics simulation application and open-source libraries are released under Apache 2.0 on GitHub, meaning you can use, modify, and distribute them freely, including for commercial purposes. SAILFISH-CFD and MESHFREE optimizations may have different terms — contact the consortium partners for specifics on those applications.

How much faster or cheaper will our simulations actually get?

Based on the project objective, OPTIMA applications running on FPGA-based HPC systems achieve more than 10x better energy-delay product than CPU-based systems and more than 3x better than GPU-based systems. This means both faster execution and lower energy consumption combined into a single efficiency metric.

Do we need specialized FPGA programming skills in-house?

OPTIMA delivered optimized programming environments specifically to lower the barrier for using FPGAs. The open-source libraries provide reusable building blocks so your developers don't need to program FPGAs from scratch. However, some familiarity with heterogeneous computing would still be beneficial.

Is there ongoing support after the project ended?

The project ended in November 2023 and planned an open-to-use HPC infrastructure supported by a sustainability body. The open-source code on GitHub remains available. For ongoing commercial support, you would need to engage directly with the consortium partners, particularly the 6 SME participants.

Consortium

Who built it

OPTIMA's consortium of 10 partners across 6 European countries (Greece, Germany, Spain, Italy, Netherlands, Switzerland) is heavily industry-weighted at 60%, with 6 of the 10 partners being SMEs. This is an SME-driven project, which means the tools were designed for companies that actually need to run simulations cost-effectively — not just for academic benchmarks. The coordinator is the Technical University of Crete (Polytechneio Kritis), a research institution, providing the scientific backbone while SMEs shaped practical requirements. The mix of 1 university, 3 research organizations, and 6 industry players suggests the outputs are grounded in real commercial needs rather than pure research curiosity.

POLYTECHNEIO KRITISCoordinator · EL
FORSCHUNGSZENTRUM JULICH GMBHparticipant · DE
APPENTRA SOLUTIONS SLparticipant · ES
ENGINSOFT SPAparticipant · IT
EXASCALE PERFORMANCE SYSTEMS - EXAPSYS IKEparticipant · EL
CYBERBOTICS SARLparticipant · CH
EREVNITIKO PANEPISTIMIAKO INSTITOUTO SYSTIMATON EPIKOINONION KAI YPOLOGISTONparticipant · EL

How to reach the team

The coordinator is Polytechneio Kritis (Technical University of Crete), Greece. SciTransfer can facilitate an introduction to the right consortium contact for your specific application domain.

Order a report on this consortium See POLYTECHNEIO KRITIS profile →

Next steps

Talk to the team behind this work.

Want to know if FPGA-accelerated computing fits your simulation workload? SciTransfer can match you with the right OPTIMA partner for a focused technical conversation.

Order a report on this topic More in Digital & ICT