extreme computing services

Accelerate your HPC production with Atos services

Our HPC service teams offer a complete portfolio of professional services, with the required competence and experience to support the design, installation, operation, management, and continuous improvement of HPC infrastructures.


Data management

Application performance

Application workflow

Data center & energy efficiency

I/O optimization


Service operation




Data Management

Managing massive data sets

Extreme data is a set of skills, experience and technologies offered by Atos to manage the complete lifecycle of your data in an optimized way. Extreme data addresses in particular data workflow analysis, HSM configuration, I/O optimization, as well as data storage, restore and movement.

More information on extreme data

Some references for Data Management



Application Performance Optimization

Reap immediate benefit from the latest HPC technologies

Atos experts can help you ensure your applications make the most of the latest HPC technologies and the bullx parallel computing capabilities.

Our expert services include:

  • Recommendations to application developers so that they can make the most efficient use of new parallel technologies (accelerators, co-processors, hybrid systems, many-core processors…)
  • Optimization of commercial  and open-source applications
  • Proof-of-concept to introduce end-users to the benefits of parallelism at the start of the project
  • I/O and network optimization
  • Ensuring applications make the most of your Bull supercomputers from day one, with the Fast Start program that addresses the porting, optimizing and configuring of your applications well ahead of system delivery.

More information on the Atos Center for Excellence in Parallel Programming

More information on I/O optimization

Some references for Application Performance Optimization



Application Workflow

Increase the reliability of your applications

Atos experts can help you manage your application workflow. Atos's best practices rely on the ITIL recommendations.

Optimizing the application workflow starts with the strategy defining the user need analysis, and spans to the design phase with the definition of SLAs, and to the transition and the operation phases. 

The transition phase includes the integration, acceptance  and  pre-production tests. The operation phase includes the incident and problem management and an on-going optimization.

More information on Service Operation


Data Center and Energy Efficiency

Bring the PUE (Power Usage Effectiveness) of your data center as low as 1.1 - or even lower

There are many ways to optimize energy consumption, from the design of a green hardware architecture, with Direct Liquid Cooling technologies for example, to the optimization of the cooling system, and including also the scheduler configuration, and the setting of energy KPIs to make your applications energy-aware.

Atos has a long-standing experience of delivering and supporting the largest data centers, and has thus acquired unique experience in all aspects of energy efficiency. Atos's range of expertise around data center efficiency includes customer site analysis, logistics requirements, standard compliance, data center integration, and maintenance and support.

For example, to optimize data center design, Atos can simulate an entire computing room, including air flows for optimum cooling, and modeled cable routes and rack locations.

More information on bullx direct liquid cooling servers

More information on mobull, the Bull data center in a container

Some references for Data Center and Energy Efficiency


I/O Optimization

Boost application performance by 30%

Optimizing the I/O to boost data flow and application performance can bring a gain in performance of up to 30%. To help you achieve this, Atos experts can run a performance audit, evaluate the workflow balance, tune the write operations, the data traffic on the controllers, optimize the data transfer, tune the file system, ensure all meta data are located on high-speed disks, and define the  storage architecture.

More information on extreme data

More information on application performance


Service Operation

Optimize production and operation costs

As a matter of efficiency and operation costs, Atos experts ensure you remain compliant with your SLAs. 

Atos teams can run your operations until the skills transfer and handover to your team. Our service operation consists in a remote monitoring coupled with event management, trouble ticketing coupled with incident management and user request management, change management with impact analysis and on-going optimization service including problem management, knowledge management and documentation. 

Some references for Service Operation




As a standard part of the installation process, Atos provides on-site, tailored, workshop training for the system administration team. Atos can also deliver on request a range of advanced training programs for HPC users and HPC developers, which will help your teams make the most of their Bull supercomputer and run the operations in an optimal way.

Proposed training courses include application development and optimization, how to port applications and what compilation options to select, how to improve performance, for example how to improve job submissions, or how to choose message queues. Some training course are dedicated to operations and system administration, OS management and tools (batch manager, file system management, network management, monitoring…). A  specific data management course focuses on data assessment.



White Paper

Atos extreme data is a set of technologies, skills and professional services dedicated to intensive data management and data analytics, in order to tackle the complete data life-cycle, from data generation to data processing, data analytics, mining and knowledge discovery, data storage and long-term preservation.

Read the White Paper, including the Meteo France use case >>

Contact an expert

Xavier Vigouroux
Xavier Vigouroux
Director of the Center for Excellence in Parallel Programming, Atos