POP Supports SMEs to get more out of their Investment in Parallel Software

Friday, March 6, 2020

High Performance Computing (HPC) is the engine that powers the digital economy. The availability of HPC computing in the cloud, along with the decline in the price of supercomputer hardware, have made HPC services accessible to a much wider range of users, especially small and medium-sized enterprises (SMEs). HPC enables the use of advanced computing models and simulations, making it possible to process large amounts of data in a short period of time. The use of HPC provides significant potential for SMEs to increase their competitiveness in the global economy, e.g. to develop new and better products and services, to reduce design costs and to shorten the time to market.

While the demand for HPC services has been increasing in all key sectors of the European economy, many European SMEs are lagging behind. These SMEs still use in-house workstations rather than far more powerful HPC resources and consequently, can only run relatively small-scale computation. This conservative choice limits their potential for innovation and damages their long-term competitiveness.

Limited knowledge of the potential benefits of HPC, lack of in-house expertise, and difficulties in raising finance for investment in HPC are the main constraints that limit the uptake of HPC services among European SMEs [1]. Support is clearly needed to ease these constraints and help SMEs to embrace HPC. Several EU-funded initiatives exist to help SMEs get on the HPC ladder and use these systems effectively. The EU Centre of Excellence Performance Optimisation and Productivity (POP) is one of these, focusing on performance optimisation and productivity. POP puts the world-class HPC expertise of eight commercial and academic partners at the disposal of SMEs.

POP helps SMEs to improve their parallel software by first identifying performance issues and then by supporting them in implementing suggested improvements. In the first phase of POP, between 2015-2018, we provided free-to-user services to thirty-four commercial organisations, where twenty-five of them were SMEs. This resulted in an average performance improvement of 2.25, i.e. on average we more than halved the time to solution. For a list of POP success stories, visit success stories. A few examples are:

3x speed improvement for zCFD, a computational fluid dynamics solver (for details, see here).
About 4x performance improvement for SCM’s Amsterdam Modeling Suite (AMS), a popular computational chemistry software (see here).
25% performance improvement for Urban Microclimate simulations (see here).
44% runtime reduction for Shearwater Reveal, a seismic processing code (see here).

POP has introduced a methodology for performance assessment and optimisation that allows productive and efficient collaboration with users across all sectors. POP uses a set of metrics to assess the performance of parallel codes. These metrics provide a standard, objective way to characterise different aspects of the performance of parallel codes. These simple measures, along with the performance data and recommendations for code improvements, are presented to users.

The second phase of POP was launched in December 2018. The EU H2020 programme has funded POP for a further three years to allow continuity of the much-needed services for HPC users in Europe. The second phase of POP has collaborated with 10 SMEs so far and continues to provide performance assessment and optimisation services, free to users, for European SMEs. To get in touch, email pop@bsc.es. To read more, visit services. To request our services, fill in this form.

[1] B. Gigler, A. Casorati and A. Verbeek, “Financing the future of supercomputing How to increase investment in high performance computing in Europe”, available at: https://www.eib.org/attachments/pj/financing_the_future_of_supercomputing_en.pdf, European Investment Bank, 2018.