NVIDIA Components, HPC

Higher Performance With Larger, Faster Memory
The NVIDIA H200 Tensor Core GPU supercharges generative AI and high-
performance computing (HPC) workloads with game-changing performance
and memory capabilities.

Based on the NVIDIA Hopper™ architecture, the NVIDIA H200 is the first GPU to
offer 141 gigabytes (GB) of HBM3e memory at 4.8 terabytes per second (TB/s)—
that’s nearly double the capacity of the NVIDIA H100 Tensor Core GPU with
1.4X more memory bandwidth. The H200’s larger and faster memory accelerates
generative AI and large language models, while advancing scientific computing for
HPC workloads with better energy efficiency and lower total cost of ownership.

Data sheet(s)

Lenovo Workstations

The ThinkStation PGX, a compact, personal AI developer workstation. The ThinkStation PGX is ideal for AI researchers and developers, data scientists, practitioners, students, and application engineers who need a purpose-built, compact, and powerful AI desktop solution that works immediately out of the box.

Built on the NVIDIA GB10 Grace Blackwell Superchip providing up to 1 PetaFlop (1000 TOPS^[1]) of AI performance, the ThinkStation PGX can tackle large generative AI models of up to 200 billion parameters. With 128 GB of coherent unified system memory^[2], developers can experiment, fine-tune, or inference the latest generation of reasoning AI models. To double down on computing power, developers can connect two ThinkStation PGX systems together to work with even larger AI models up to 405 billion parameters.

The ThinkStation PGX comes preconfigured with the NVIDIA DGX OS,  and the NVIDIA AI software stack, along with familiar tools and frameworks like PyTorch and Jupyter. Developers can instantly prototype, fine-tune, and inference large AI models from the desktop and seamlessly deploy to the data center or cloud.

IBM Storage

IBM Storage

IBM Storage Scale System 3500

IBM Storage Scale System 3500 is for customers who require an entry-level or midlevel system for less extreme product environments. It has many of the same capabilities as the Storage Scale System 6000, with differences mainly in size and capacity.

IBM Storage

Preserve and protect data with sustainable, cyber-resilient, ultra-high-density tape storage

Next-generation sustainable ultra-high-density storage, which builds on more than two decades of storage innovation. IBM® Diamondback delivers high-capacity, long-term data storage with the air-gapped cyber resiliency only tape can provide.

Organisations face challenges in managing vast amounts of data, especially those with hundreds of petabytes, in fields like machine learning, autonomous driving, pharmaceutical research, and astrophysics.

​They require reliable storage, backup, and recovery systems to protect against natural disasters and cyberattacks.

KEY BENEFITS OF DIAMONDBACK:

Ultra high-density storage
IBM Diamondback delivers ultra high-data density, with up to 46.4 PB2 of native data in a single eight-square-foot library using LTO Ultrium 10 cartridges.

Massive capacity
Using standard 2.5:1 compression, each IBM Diamondback Tape Library can store up to 116.1 PB, which provides the high-capacity organisations need to retain their growing volume of data.

Enhanced security
IBM Diamondback features tape-drive encryption, automatic data-path failover, dual robotic grippers, and WORM media support to improve management and help reduce risk in operations.

Enhanced cyber resilience
Because it’s physically separated from an organisation’s primary storage systems, IBM Diamondback Tape Library provides an “air gap” that helps isolate archived data and improves cyber resilience.

Data sheet(s)

IBM Storage

The IBM® TS3500 tape library is a highly scalable, stand-alone device that provides reliable, automated tape handling and storage for both mainframe and Open Systems environments.

An individual library, or library string, consists of one base frame and up to 15 expansion frames and can include up to 192 tape drives and more than 20 000 tape cartridges. The flexibility to connect multiple library strings to create a shuttle complex greatly increases opportunities for growth, as well as the maximum cartridge capacity. In addition to scalability, the TS3500 tape library offers the following enhancements:

• Enhanced data accessibility through dual accessors that increase speed and provide failover protection
• Enhanced data security through support for tape drive encryption and write-once-read-many (WORM) cartridges
• Increased storage capacity with high-density (HD) frames that greatly increase capacity without requiring more floor space
• Increased growth flexibility with shuttle technology, even in physically constrained environments
• Greatly increased cartridge capacity in a shuttle complex without the need for additional tape drives.

The TS3500 tape library comes with several tape drive, frame model, and feature options to meet your specific needs. Some additional features of the TS3500 tape library are listed here:

• Ability to attach multiple simultaneous heterogeneous servers
• Remote management using a web browser or the TS3500 Command Line Interface program
• Remote monitoring using Simple Network Management Protocol (SNMP)
• Multipath architecture
• Drive/media exception reporting
• In-depth reporting using the Tape System Reporter
• Host-based path failover
• Up to 224 I/O slots (16 I/O slots standard)

OCF Hosting

OCF partners with hosting provider DataVita who offer high density, sustainable hosting ideal for HPC Applications.

DataVita is a prominent provider of data center and multi-cloud solutions, with a strong emphasis on supporting high-power, complex workloads, including those for Artificial Intelligence (AI) and High-Performance Computing (HPC). The company's vision is to become the leading co-location and cloud services provider in Scotland and to expand across the wider UK market. Their infrastructure is designed to handle demanding environments, ensuring the resilience, security, and efficiency necessary for critical HPC operations. DataVita is a sustainable partner, claiming that its clients can significantly reduce their digital emissions, aligning with modern corporate responsibility goals.

The company's data centers are built for maximum performance and reliability. DataVita has achieved Uptime Institute Tier III accreditation for both the design and construction of its facilities, guaranteeing high availability and fault tolerance. They also back their services with a 100% uptime service level agreement (SLA), providing assurance for partners and clients running mission-critical applications. This robust foundation is essential for HPC workloads that require consistent and uninterrupted access to computing resources.

DataVita's commitment to scalability and strategic location is evident in their data center portfolio. Their flagship facility, DV1 in Lanarkshire, was the first in the UK to receive Tier III certification, while their new DV2 data center is centrally located in Glasgow. With a total of five sites, including two live and three in development, DataVita is well-positioned to meet the growing demands for high-density, low-latency computing environments. This network of facilities allows for flexible co-location and cloud deployment models, which are crucial for distributed and complex HPC projects.

By focusing on high-power workloads and combining efficient, resilient infrastructure with a sustainable approach, DataVita presents a compelling value proposition for partners in the HPC sector. Their ability to handle diverse and complex computing environments—from legacy systems to high-power AI and low-latency applications—makes them a suitable collaborator for organizations seeking to optimize their computing infrastructure. Their services offer significant cost savings, with clients potentially saving up to 60%, and their investment in renewable energy underscores their forward-thinking approach to the future of data and computing.

ddn Storage

The AI400X2 is DDN’s highest-performing system purpose built for large scale AI training and used in many of the largest AI clusters as part of a NVIDIA DGX SuperPOD.

The systems compact low power all NVMe design combined with DDN’s enterprise EXA Lustre-based filesystem and native InfiniBand support make this box well suited to the most IO challenging applications.

Data sheet(s)