Today’s increase in terms of quantity and complexity of data leads to an exponential need of computing power. Atos next-gen hybrid supercomputing platform BullSequana XH3000 delivers unprecedented computing power at any scale, up to an exascale-class system. With its open and standardized architecture, coupled with AI-empowered smart software suites, Atos BullSequana XH3000 is ready for accelerated hybrid computing and enables customers to tackle the most pressing challenges in the 21st century.
The accelerated hybrid HPC platform
Combining cutting-edge processing technologies, Atos 4th generation (DLC) Direct Liquid Cooling technology, and an architecture that is flexible, dense and secured by design, BullSequana XH3000, delivers both unprecedented performance and unrivaled efficiency.
With Open Sequana program, Atos allows for 3rd party technology ecosystem participation and delivers to customers a future-proof platform ready for quantum accelerators.
BullSequana XH3000
We are never at rest.
Our curiosity pushes us out.
Constantly questioning,
We explore and wonder at phenomena
as rational as they are magical.
Because there is always an explanation.
When we shape our questions,
We open new doors.
When we explore our ideas
we decode new paths…
Our joint discovery is illuminating,
The perspective, that is the feat.
Any scale & Exascale
All-in-one rack design & seamless scalability
Industry-leading density
Maximized node performance & optimized footprint
Unrivaled efficiency
The 4th Gen Direct Liquid Cooling & AI augmented software
Secure by design
Hardware-anchored security boot a threat anticipation
Unprecedented flexibility
Standardized slots & Open Sequana embracing 3rd party innovations
Atos supports the arrival of the 4th Gen AMD EPYC processors, instrumental for our next-gen hybrid HPC, BullSequana XH3000
In today’s rapidly changing world, collaboration/partnership is essential for accelerated innovations. Partnering with trend-setters enables value creation at scale and speed. Atos is thrilled to welcome the arrival of the 4th Gen AMD EPYC™ processors which will feature the Atos BullSequana XH3000 to continuously raise the bar yet again for workload performance and a new level of efficiency.
What do the industry analysts say
“Today’s high-performance applications have expanded to include analytics and artificial intelligence, fueling new types of data-driven discovery. The next generation of supercomputing has to scale multidimensionally to incorporate these new types of workloads. With hybrid computing capabilities and low carbon footprint, the Atos Bullsequana XH3000 is designed to meet these challenges, across a broad range of commercial engineering and analytics, up to Exascale and beyond.”
Addison Snell
CEO, Intersect360 Research
“High-performance computing has made enormous contributions to science, engineering and the quality of human life and is a cornerstone of innovation and scientific progress“ said Earl Joseph, CEO at Hyperion. “As the leading European HPC vendor, Atos is well positioned to continue advancing research and innovation with its new supercomputer, and to bring Exascale class systems to its customers. With its new BullSequana XH3000 supercomputer, hybrid computing ecosystem integrating AI and quantum, and its strong roadmap for Exascale, Atos demonstrates a clear, comprehensive vision of the future of scientific simulation, while taking into account the importance of decarbonization.”
Earl Joseph
CEO, Hyperion Research
Andy Grant
VP Global Sales, HPC, AI and Quantum, Atos
Follow or contact Andy:
Julien Camiade
Head of Hardware Portfolio & Solutions, HPC, AI and Quantum
Follow or contact Julien:
BullSequana X800: at the confluence of HPC and Big Data
Featuring an architecture that supports an impressive number of processing cores and an exceptional memory capacity, the BullSequana X800 servers are engineered for high performance and in-memory computing. They are ideally suited to support massive in-memory databases, pre-processing, post-processing, and visualization. They are among the most scalable and modular ultra-high memory servers on the market, starting with 2 CPUs and 24 memory DIMMs, and going up to 32 CPUs with 384 memory DIMMs in 16 interconnected boxes – with a linear performance scaling.
–
A powerful and modular range
The BullSequana X800 series consists of five complementary models, assembling 1 to 16 2-socket compute boxes thanks to two types of interconnects:
- A glueless interconnect is used for 1, 2 or 4 compute boxes, allowing to form a glue-less configuration with up to 8 sockets.
- For larger configurations, a UNC (UPI Node Controller) stacks 8 to 16 boxes.
This distinctive Atos technology makes it possible to form 16- or 32-socket servers.
A Connecting Box assembles compute boxes with no apparent cabling, and makes it possible to easily add compute and memory resources. This helps adjust memory and compute resources to application needs, and facilitates the scaling of Big Data applications. For more agility in configuring, BullSequana X800 servers feature hot plug I/O blades, disks, fans and PSUs.
–
Key innovative components
Each compute box is a 2U 2-socket rack server and contains:
- up to 24 memory DIMMs, i.e. a total of up to 3 TB per compute module when using 128GB DIMMs – and ready for higher capacities with future memory technology;
- hot plug I/O blades that offer up to 5x PCIe Gen3 slots, with capability for a disk controller in charge of the front-end disks;
- a number of options to customize the system to match application requirements: internal disks, NVMe drives or GPUs.
–
The performance of leading-edge Intel® Xeon® Scalable processors
The BullSequana X800 embarks Intel® Xeon® processors Scalable family, with a large choice of models for the best fit with your applications in terms of:
- frequency,
- number of cores, or
- power consumption.
–
Designed for operational efficiency and maintainability
The BullSequana X800 focuses on serviceability and operational efficiency. It has all required features for reliability, availability and quality of service:
- redundancy of critical components;
- hot-pluggable critical components – the innovative blade form-factor of I/O modules enables hot-plugging;
- patented mechanics for easy replacement of components: the disks, fans and PSUs can be extracted without rack decabling and without stopping running applications;
- server partitioning;
- internal regulation and alert systems;
- dedicated maintenance processor;
- power-conscious design: with the Gen2 Active/Passive Power supply, energy consumption is reduced by 40%.
Press release
BullSequana X400 series
Truly designed for any HPC workload
The BullSequana X400 series is a comprehensive family of rack-mounted servers designed for High Performance Computing and offering an optimum balance between density, efficiency, and cost.
Cutting-edge technologies in a modular solution
By integrating the latest industry leading CPU/GPU architecture (Intel® Xeon® Scalable processors, AMD EPYC™ processors, Nvidia® A100 GPU) in modular servers, Atos’ BullSequana X400 series is the perfect compromise between high performance and versatility within a HPC cluster.
With a broad range of nodes (compute, service, GPU accelerator and graphic) BullSequana X400 series allows you to address any workloads and to scale as you grow. Flexible interconnect solutions (InfiniBand EDR, HDR, High Speed Ethernet) and modular software solutions can handle the most demanding HPC system, data visualization and rendering workloads.
Protect your investment
Fast Start program & CoE in Advanced Computing
When choosing Atos, a dedicated HPC player, you choose the long-standing expertise of a trusted partner.
We give you the keys to execute your workloads efficiently the quickest possible way. From day one, you benefit from the highly professional services capability, thanks to Fast Start program with Atos experts. Moreover, Atos Center of Excellence in Advanced Computing collaborates with you to release the maximum potential of your system. On-time security fixes delivery, high quality maintenance and long-time support under SLA to ensure maximum uptime, will secure your investment during the whole lifecycle of your system.
Optimize your TCO
BullSequana X400 family is designed to be cost-effective without leaving performances out.
Top efficient air-cooling system supporting highest TDP per node in ultra-dense rackmounted server delivers a tailored flexible HPC solution to market. It is the optimum balance between density, performance and cost.
A broad range of cutting-edge servers
AMD EPYC™ processors
BullSequana X440 A5 2U4N1S – 2U4N2S
Compute node
Ultra-dense server
Integrate 4 dual or single socket nodes within a 2U standard rack chassis
Delivered with latest generation AMD EPYC™ series processors and fastest Mellanox InfiniBand or Ethernet network adapters
Offer even more performance with efficient air cooling
Ideal modular compute nodes server within a HPC cluster
BullSequana X430 A5 2U1N1S – 2U1N2S
Service node
Advanced connectivity features and extended storage options
Efficient service node with high flexibility in 2U form factor, available in single and dual socket
Wide high-speed connectivity range (InfiniBand, Ethernet and fiber channels) thanks to its numerous PCIe Gen4 expansion slots available
Large set of storage capability (HDD, SSD, NVMe, M.2)
BullSequana X410 A5 2U1N1S 4GPU – 2U1N2S 8GPU
GPU Accelerated node
Performance and power efficiency
Able to integrate up to eight GPU (Nvidia® V100s, A100 or AMD® Radeon Instinct) within 2U, interconnected via PCI direct connect
Feature single or dual AMD EPYC™ series processor
Ideal for HPC applications such as real time analytics, programs in scientific simulation visualization…
BullSequana X450 A5 2U1N2S 3GPU
Graphics node
Committed to versatility with high performance
Flexible dual socket server
Optimum balance between connectivity, storage and GPU support
Optimized to support up to three GPUs and the latest NVIDIA® Tesla® V100s, A100, Quadro RTX or Radeon Instinct
Handle most demanding data visualization and rendering workloads
BullSequana X410 A5 2U1N2S 4GPU SXM
Most Advanced GPUs with support of NVIDIA HGX A100 4-GPU in a 2U air cooled form factor.
Ideal for the Convergence of Simulation, Data Analytics, and AI to accelerate scientific progress.
Allows most complex HPC problems to be solved without being limited by compute capability.
GPU processors
GPU Accelerated compute node
BullSequana X410 A5 2U1N1S 4GPU – 2U1N2S 8GPU
Performance and power efficiency
A complete and ultra dense GPU-based solution
Able to integrate up to eight GPU (Nvidia® V100s, A100 or AMD® Radeon Instinct) within 2U, interconnected via PCI direct connect
Feature two AMD EPYC™ series processors
Ideal for HPC applications such as real time analytics, programs in scientific simulation visualization…
BullSequana X410 A5 2U1N2S 4GPU SXM
Most Advanced GPUs with support of NVIDIA HGX A100 4-GPU in a 2U air cooled form factor.
Ideal for the Convergence of Simulation, Data Analytics, and AI to accelerate scientific progress.
Allows most complex HPC problems to be solved without being limited by compute capability.
Graphics node
BullSequana X450 A5 2U1N2S 3GPU
Committed to versatility with high performance
Flexible dual socket server
Optimum balance between connectivity, storage and GPU support
Optimized to support up to three GPUs and the latest Nvidia® Tesla™ V100s, Quadro RTX graphic cards
Can handle the most demanding HPC, data visualization and rendering workloads
HPC Software Suites – introduce a new approach to eXtreme computing software solutions.
They are scalable, open, and robust software suites that meet the requirements of the most challenging high-performance computing (HPC) environments, with high security standards.
Suited for small to Exascale configurations
HPC Software Suites are the result of Atos’ long experience in deploying large-scale supercomputers, combined with the continued efforts in Research & Development.
HPC software suites are designed to address every HPC need, whether it is for a small system with just a few cores or for supercomputers with tens of thousands of cores. They are also designed to reach performance targets of the most powerful Top500 systems. The main goal is to provide a global high performance supercomputing environment that includes:
- The Smart Management Center: install, configurate, and operate HPC clusters at all scales
- The Smart Software Suites (optional) to understand, control and optimize energy, data, and performances
- The adds-on AI capabilities: transverse applications and broader business applications leveraging AI to manage data assimilation and security challenges
HPC Software Suites are a further step towards Exascale computing and decarbonization.
The software suites for your supercomputer and your applications
This modular software solution that manages supercomputer environments, including data and applications is based on an enterprise-class Linux distribution: Red Hat Enterprise Linux 7 and 8.
It provides a robust, scalable, and flexible solution that adapts easily to any customer needs.
All critical services are configured to ensure high availability.
All software suites are based on best-of-breed open-source and ISV software.
The selected open- source components are enhanced by Atos’ added-value features.
ISV components can be added or can fully replace default components, depending on user needs.
REST APIs are available in all those software suites for seamless integration in any existing customer environment.
By design, the modular software approach is allowing any ISV, open-source, or home-grown software and tools to run seamlessly.
Atos services to take full advantage of supercomputer suites
As a leading provider of end-to-end Extreme Computing solutions, Atos provides consultancy services to help its customers take full advantage of the benefits of those software suites throughout their projects: design, deployment, and operations.
Customers can benefit from Atos’ highly professional worldwide support offer that includes traditional maintenance and upgrade services as well as dedicated, customized, and proactive services.
Related resources
Press release
November 19, 2020
Atos announces new HPC Software Suites to optimize performance and reduce energy consumption
Shifting to a new scale
Shifting to a new scale
Exascale entails an explosion of performance, of the number of nodes/cores, of data volume and data movement. At such a scale, optimizing the network that is the backbone of the system becomes a major contributor to global performance.
The interconnect is going to be a key enabling technology for exascale systems. This is why one of the cornerstones of Atos’ eXascale program is the development of our own new-generation interconnect.
The BullSequana eXascale Interconnect V2 or BXI V2 introduces a paradigm shift in terms of performance, scalability, efficiency, reliability and quality of service for extreme workloads.
BXI V2 components
The BXI V2 fabric relies on two types of ASICs as its building blocks, a Network Interface Controller (NIC) and a switch, and comes with its complete software suite. BXI V2 switches are managed through a distributed and out-of-band fabric management suite allowing to scale up to 64K nodes. Out-of-band management eliminates any interference of the management traffic with the applications.
Interconnect for hybrid platforms
While BXI V1 was dedicated to BullSequana X1000 systems, BXI V2 is available on all BullSequana X platforms: it integrates the latest industry leading technologies within all BullSequana X servers, from BullSequana X400 mid-range series to BullSequana XH2000 high-end systems. Moreover, BXI V2 Network Interface Controller supports AMD, and ARM compute nodes, in addition to Intel ones.
The highest level of communication performance for HPC applications
BXI V2’s main feature is full hardware offload of communications. Indeed, by offloading to the hardware components, BXI V2 communications do not interfere with computations.
Offload boosts the communication performance of HPC applications, enabling CPUs to be fully dedicated to computational tasks while communications are independently managed by BXI V2.
As a result, BXI V2 can deliver high communication performance considering bandwidth, latencies, or message rates.
Fabric Management
Advanced Fabric Management (AFM) is responsible for monitoring and routing a BXI V2 fabric while IBMS handles Device Management and Performance Management tasks.
With a new patented routing algorithm, AFM boosts the performance of BXI V2. BXI V2 allows a wide range of topologies: Generalized Fat-Tree, and DragonFly +.
AFM and IBMS software are fully compliant within SuperComputer Suite, a scalable, and robust software suite that meets the requirements including the need for enhanced security of even the most challenging high-performance computing environments.
A wide ecosystem available for application environment
The application environment of BXI has been significantly developed between V1 and V2. BXI V2 now enables Intel MPI library support thanks to an OFI provider.
Moreover, BXI V2 brings new features such as:
- GPU direct
- NVMEoF
- Virtualization (up to 16VMs)
BullSequana Xstor
A single family of storage appliances
Today Machine Learning and Data Analytics are creating many of the most challenging HPC workloads as we move inexorably towards the exascale era. Atos actively participates in initiatives to strengthen the industrial control of digital simulation and Big Data technologies.
At Atos we believe that HPC users are just beginning to feel the possibilities of Hybrid Computing. After developing BullSequana XH2000 the first fully hybrid computing solution, we introduce our hybrid storage solution. The goal is to support our customers as a trusted partner, in addition to meeting their specific needs.
The BullSequana Xstor from Atos is a new range of modular storage appliances designed for HPC current and future needs. We developed this solution with that new paradigm in mind, focusing on Performance, Modularity, Scalability and Cost Optimization to deliver our customers a new range of capabilities.
Investment of Atos in Storage
Atos is a leader in digital services focusing on developing technology products and software. We actively participate in initiatives to strengthen the industrial control of digital simulation and Big Data technologies. Atos long-time experience in storage is a the perfect base for building and supporting solutions required by data centric architectures.
BullSequana Xstor Main Benefits
High level of Performances
Xstor is a storage appliance dedicated to HPC with high performance Storage features, including high performance Interconnect and the Highest Density.
Scale as You Grow
Scale-out capabilities: Match your storage needs with the highly scalable building block architecture.
Scale-up capabilities: Using JBODs extensions.
Modular Architecture
BullSequana Xstor’s architecture has been designed with modularity in mind.
Customers are either able to focus on optimized performance (bandwith) or large storage capacity.
BullSequana Xstor Use Cases
Parallel File System
BullSequana Xstor using Spectrum Scale is the perfect solution for a parallel file system performance dedicated to performance, resiliency and scalability.
Data Lake & AI
Using CEPH-based Software Defined Storage, handle all your data and push even further your Data Lake & AI applications in an HPC environment.
DataTiering
An HPSS data tiering solution using BullSequana Xstor for disk caching is fully optimized for archiving huge data volume for a long time with high level of performance.
.
