Yesterday, Atos announced three new HPC cloud servers that have the computational power to tackle trends in the engineering industry such as big data, artificial intelligence and machine learning (AI/ML). The first option is a hybrid server named BullSequana SH; the other two are edge servers named BullSequana EXR and BullSequana EXD.
Stefan Kero, Atos’ Head of Advanced Computing in Central Eastern Europe, said: "High-performance computers are among the key technologies of our time and have become indispensable [not only in the] sciences, medical research and weather forecasting, but also in industry, [especially] digital product development. With HPC, industrial players can accelerate the construction of digital prototypes – which can be run through quickly and as often as desired in a multitude of variants – therefore [gaining] flexibility and innovative strength.”
According to Statista, global data creation is expected to grow to 180 zettabytes by 2025. Atos argues that most computational infrastructure options aren’t designed to handle data creation at this scale. Meanwhile, the Uptime Institute Global Data Center Survey 2022 found that 63 percent of organizations are not putting their mission-critical information onto the public cloud. As a result, Atos argues that these organizations will need to rely on hybrid and edge servers—like the ones that were just announced.
“We are excited to today unveil our next generation of enterprise and edge servers,” said Emmanuel Le Roux, SVP, Group SVP Global Head of Advanced Computing & AI, Atos. “This is a key milestone for [Atos’] computing business as our full range of servers is now entirely renewed, together with last year’s BullSequana XH3000 launch. We have leveraged our recognized manufacturing and R&D expertise to develop the next generation of our edge and business computing [server] lines to help our clients tackle the challenges of fast-growing data, security and sovereignty."
All three of the new servers are optimized for flexibility, security and a low carbon footprint. Atos notes that they can run various workloads and applications with low latency and high energy efficiency due to improved compute density. Their security is based on Atos Root of Trust and Chain-of-Trust systems.
BullSequana SH for On-premises and Public Cloud Computations
The BullSequana SH server is based on the 4th Gen Intel Xeon Scalable processor. Thus, it can scale with 2-socket increments up to 32 sockets working as one server. The aim is to offer users scalability without over-provisioning, and to make software more efficient without increasing energy footprints.
The server’s DDR5 memory will offer customers fast response times during demanding computations. The server also has more memory per system and uses the Intel Optane Persistent Memory 300 Series.
In summary, BullSequana SH is marketed as an accelerator to digital transformations. It helps users run large computational workloads, business-critical applications and in-memory applications such as database crunching and analytics. The server is also able to offer virtualized and hyperconverged infrastructures for hybrid computing, AI, deep learning and ML.
BullSequana EX for Edge to Cloud AI Computations
As for the BullSequana EX series, it is designed to work with data sets that are too large or business critical to move to the cloud. The EX series focuses on edge-to-cloud performance that improves availability with limited latency while working with complex and large data sets.
Those looking to use a far-edged desktop with ‘rackable’ options are suggested to go with BullSequana EXD, while those interested in an edge datacenter with hyperconverged infrastructure are recommended to go with BullSequana EXR. Either way, users get interchangeable CPUs (including 4th Gen Intel Xeon Scalable processors), and GPUs ranging from NVIDIA A2 Tensor Core GPU to the NVIDIA L40 GPU.
Atos says that the BullSequana EX servers will help engineers implement computer vision use cases such as AI inference, smart glasses, AR/VR and more.
Success for BullSequana in the Engineering Industry
Judging by Atos’ BullSequana XH2000 server announcement with Australian automotive supplier AVL List, the company has proven its mettle to the engineering community. AVL provides mobility companies assistance with product development, simulation and testing of powertrain, workpiece, drive unit and large engine designs. Thanks to Atos, AVL can provide these complex computational services at half the time while consuming less power.
"AVL’s passion is innovation and Atos’ supercomputer meets that requirement in terms of performance and energy-efficiency,” said Günter Bachler, Senior IT System Administrator, AVL List GmbH. “The direct liquid cooling (DLC) technology of the XH2000 cluster has opened up an unprecedented performance range, with minimal energy requirements and a green footprint, because 95 percent of the waste heat is not released into the environment but is put to further use via the cooling water circuit."
The computations AVL provides—including finite element analysis, structural optimizations, crash/battery cooling/combustion simulations and computational fluid dynamics—have translational applications throughout the engineering industry. As a result, engineers that use these tools should expect good performance when using BullSequana servers when designing any product or process.