Skip to main content

Public Procurement of Innovative Solutions for High-Performance Computing

Periodic Reporting for period 2 - PPI4HPC (Public Procurement of Innovative Solutions for High-Performance Computing)

Reporting period: 2018-04-01 to 2020-03-31

In its communication on the “European Cloud Initiative” the European Commission in April 2016 stressed the need for building an European Data Infrastructure including high-end supercomputers, which will around 2022 reach the exascale performance level. In this context, a group of leading European supercomputing centres proposed the formation of a buyers group to execute a joint Public Procurement of Innovative Solutions (PPI) in the area of high-performance computing (HPC). Including the co-funding by the European Commission (EC) allowing a significant enhancement of the planned pre-exascale HPC infrastructure from 2018 on, the total investment was planned to be about € 73 million.

The involved HPC centres, namely BSC, CEA/GENCI, CINECA and JUELICH, have a strong track record in providing supercomputing resources at European level. The project participants are founding members of PRACE and have made Tier-0 systems available through PRACE. This joint initiative was designed to create benefits in multiple respects:
▪ More advanced supercomputing resources efficiently exploitable for science and engineering applications in Europe within PRACE, the pan-European HPC infrastructure, as selected, relevant applications played an important role in guiding this procurement process.
▪ R&D on HPC architectures and technologies in Europe strengthened as suitable incentives was provided by this joint procurement process.
▪ A greater weight and more impact on the design of the solutions according to the needs of scientists and engineers in Europe thanks to the coordinated approach.

The participants worked together on coordinated roadmaps for providing HPC resources optimised to the needs of European scientists and engineers. The final decision on which innovative solutions are procured at the different sites were made following these roadmaps, but remained a decision of the individual sites.

The proposed concept for implementing this PPI starts with a jointly executed market consultation. This was also the first public step, where the participants present themselves as a buyers group pursuing common interests on the market of supercomputing solutions. Taking the outcomes of the market consultation in account, common requirements and evaluation criteria were jointly formulated. These were complemented by site specific requirements and evaluation criteria. Final procurement contracts were designed to be awarded within four different lots. This effort aims at the deployment of four different systems based on a variety of innovative solutions, which provide a much broader range of capabilities as what would be achievable through a single solution.
During the initial phase the project focused on analysing common needs and identifying common technical goals, objectives and requirements. At the same time the public procurers needed to define and setup a framework for executing a joint procurement with all its legal implications. Results of this initial working phase were presented to interested suppliers during the market consultation phase. The fruitful results of the market consultation phase were taken into account for producing a tendering package, which was published in May 2018.

The joint tender was organised in four lots, one for each of the involved HPC sites. For each of the lots up to five suppliers requested participation in the competitive dialogue. The qualified candidates were jointly selected on a formal basis before the procurement procedure for each of the lots proceeded at local level.

The procedure for each of the lots have been successfully completed in early 2020 by awarding the last contract. At one site deployment of the procured system has been completed while for the others the installation is ongoing or in preparation:
* BSC is deploying a new high-performance storage infrastructure that is able to scale to more than 100 PByte as a single file space or file system
* CINECA is deploying a supercomputer optimised for compute-intensive applications with a peak performance of 32 PFlop/s
* CEA/GENCI deployed a versatile supercomputer with a peak performance of 22 Pflop/s supporting highly scalable computing applications as well as high-performance data analytics and machine learning applications
* JSC will deploy a supercomputer with a peak performance of 16 PFlop/s that is optimised for data-intensive applications and therefore features a large memory and storage capacity

Based on a set of common and site-specific requirements, suppliers were specifically asked to integrate new innovations into the offered solutions. From the larger list of innovative features, we would like to highlight the following:
* Solutions for measuring and steering energy consumption to allow for dynamic optimisation of hardware parameters such that energy is reduced with minimal impact on performance
* Integration of water-cooled power supplies that improves Power Usage Effectiveness (PUE) by reducing the energy needed for cooling the system
* New solutions for integrating non-volatile memory technologies into HPC systems to enable much faster and more flexible access to storage
* Innovative software components that facilitate easy integration of tape technology in a way that data is transferred to tape transparently to the user
The PPI4HPC project was the first project to perform a joint procurement of innovative HPC solutions involving public procurers from 4 different EU member states. The public procurers formed a buyers group and agreed on a procurement procedure that was executed with GENCI from France being in the role of a lead procurer. The effort was based on the identification of a set of common challenges towards future exascale computing capabilities. The most prominent common technical challenge was to significantly increase the performance of the systems compared to currently installed supercomputers, while keeping power consumption constant or allow only for moderate increases. A further common challenge was to improve the versatility of future supercomputers to enable new types of applications involving processing of extreme-scale data volumes, enable data analytics through novel methods like deep learning, enable in-situ remote visualisation, computational steering. To address these common needs a set of common requirements, including requirements in terms of innovation, had been formulated, which were the basis for all lots.

The joint procurement executed by PPI4HPC is expected to have the following impact:
* New science results, in particular in areas related to grand challenges like climate research, energy research, new materials and health with the upgrade (about 70 PFlop/s) of the European distributed HPC infrastructure (PRACE) according to the needs of scientists and engineers
* Improved energy efficiency and therefore energy savings as all procured solutions implement innovation in this area
* Contributing to the Joint Undertaking EuroHPC as the tendering documents of the project served as a template for the procurement procedures that are currently being implemented
Today's flagship computers at the PPI4HPC supercomputing centres