Lenovo Servers

The Lenovo ThinkSystem SR780a V3 is a powerful liquid-cooled 5U server that features two 5th Gen Intel Xeon Scalable processors and eight high-performance GPUs. This server is the ultimate Generative AI server, with advanced GPU-to-GPU communications and high-speed PCIe 5.0 connectivity between the GPUs, CPUs, and networking. It was designed fully in-house by Lenovo from ground up to provide best-in-class modularity, thermal performance and reliability.

This product guide provides essential pre-sales information to understand the ThinkSystem SR780a V3 server, its key features and specifications, components and options, and configuration guidelines. This guide is intended for technical specialists, sales specialists, sales engineers, IT architects, and other IT professionals who want to learn more about the SR780a V3 and consider its use in IT solutions.

Data sheet(s)

OCF HPC, Infrastructure, Software

OCF Steel transforms standard enterprise servers into HPC clusters—providing an integrated software platform, proven methodology, and ongoing support to build and manage demanding HPC and AI environments with confidence.

From initial deployment through daily operations to end-of-life migration, OCF Steel encapsulates decades of HPC expertise into proven best practices that reduce risk, save time, and lower costs. It's the same platform OCF, as recognised HPC experts, uses to build and manage customer environments.

Core capabilities include bare metal provisioning, configuration and orchestration, shared storage, robust security and authentication, workload scheduling, remote user access, real-time monitoring with intelligent alerting, and detailed usage reporting.

What you get

OCF Steel provides the wherewithal to successfully build and manage HPC and AI environments, providing:

• An approved and validated bundle of respected open-source solutions
• OCF’s custom integration code for seamless operation
• Step-by-step best-practice implementation and operational guidance
• Support, including updates, with access to specialist advice
• Optional, managed service with OCF taking care of day-to-day management.

Data sheet(s)

NVIDIA Workstations

The NVIDIA DGX Spark personal AI supercomputer enables developers to prototype, fine-tune and inference models on their desktop and deploy them locally or seamlessly deploy at scale to the data center or cloud.

Powered by the new GB10 Superchip based on the Grace Blackwell architecture, DGX Spark can tackle large generative AI models of up to 200B parameters.

The DGX Spark is ideal for AI developers, researchers, data scientists, and students who need a compact and powerful AI solution.

Data sheet(s)

OCF Software

SLURM (Simple Linux Utility for Resource Management) is a powerful open-source workload manager and is a key part of OCF Steel Stack HPC cluster suite.

When you have more user jobs than compute cores to run them at once, SLURM ensures each gets the opportunity to run. It is highly customisable, and our in-house team of expert engineers can configure it to fairly manage even the most diverse workloads.

It is well suited to both the largest and smallest HPC clusters. Its network-aware scheduling with an efficient back-fill algorithm can drive high utilisation across many nodes. Whilst its “free” nature is a good fit for smaller clusters where budget spent on more hardware generally improves throughput.

We can also offer commercial software support for SLURM with our partnership with SchedMD.

Altair HPC

PBS Pro is among the industry leaders in workload management and a great fit for commercial research environments where downtime needs to be minimised and security is of the upmost importance.

It is well suited to engineering workloads, with integration with many widely used commercial CAD and simulation applications, and is also used on many of the world's largest supercomputers due to its superb reliability.

It is feature rich, offering more enhanced reporting capabilities than open-source alternatives, and has a long-established track record of delivering results in secure sites.

Data sheet(s)

Lenovo Workstations

The Lenovo ThinkStation P8 is a high-performance workstation designed to tackle the most intensive professional tasks, setting itself apart with a focus on AMD's powerful **Ryzen Threadripper PRO 7000 WX-Series processors**. This single-socket platform can be configured with a staggering number of cores, going up to 96 cores and 192 threads, which provides exceptional power for heavily multi-threaded applications like 3D rendering, simulation, and data science. The P8 also offers an 8-channel memory architecture, supporting up to 1TB of DDR5-4800 ECC memory, ensuring a massive memory bandwidth for demanding datasets. In terms of graphics, it can be equipped with up to three high-end, double-width professional GPUs, such as the NVIDIA RTX 6000 Ada Generation, making it a formidable tool for GPU-accelerated workflows.

The design of the ThinkStation P8 is a collaboration with Aston Martin, resulting in a chassis that is both aesthetically striking and highly functional. The "Storm Grey" and red accents, along with the distinctive front grille, are not just for style; the grille is engineered for maximum airflow to keep the powerful components cool. The chassis features a modular, tool-less design with red touchpoints to indicate removable parts, making it exceptionally easy to service and upgrade. The system also includes an advanced thermal design with a large dual-fan heatsink for the CPU and dedicated fans for the memory modules, ensuring the workstation can sustain peak performance during prolonged, heavy workloads without overheating.

Beyond its raw power, the P8 is built for enterprise-level reliability and security. It is rack-optimized with a 4U design and optional rail mount kit, allowing for flexible deployment in either a desktop or data center environment. An optional Baseboard Management Controller (BMC) card provides IT managers with full remote management capabilities. The workstation also includes a suite of security features, such as a Trusted Platform Module (TPM) 2.0 and a self-healing BIOS, to protect data and the system itself. With its combination of a powerful AMD platform, thoughtful engineering, and robust enterprise features, the ThinkStation P8 is a top-tier solution for professionals who need uncompromising performance in a single-socket workstation.

Data sheet(s)

Lenovo Workstations

The ThinkStation P620 was the first AMD Ryzen™ Threadripper™ PRO powered workstation, delivering game-changing power and speed with the largest core count in a professional workstation: up to 64 cores and 4.5GHz. Perfect for those who demand more power for multi-threaded applications, the P620 combines Lenovo’s legendary reliability, experience and customer-driven innovation with professional manageability and enterprise-class support.

Purpose built with the speed and configurability you crave, this powerhouse workstation offers PCIe® Gen4 for increased storage speed and expandable memory, as well as unmatched NVIDIA® graphics support, and 10Gb of built-in ethernet—a new standard in connectivity.

NVIDIA GPUs, HPC

An Order-of-Magnitude Leap for Accelerated Computing
The NVIDIA H100 Tensor Core GPU delivers exceptional performance, scalability, and security for every workload. H100 uses breakthrough innovations based on the NVIDIA Hopper™ architecture to deliver industry-leading conversational AI, speeding up large language models by 30X.

Securely Accelerate Workloads From Enterprise to Exascale

H100 features fourth-generation Tensor Cores and a Transformer Engine with FP8 precision that provides up to 4X faster training over the prior generation for GPT-3 (175B) models. For high-performance computing (HPC) applications, H100 triples the floating-point operations per second (FLOPS) of double-precision Tensor Cores, delivering 60 teraflops of FP64 computing for HPC while also featuring dynamic programming (DPX) instructions to deliver up to 7X higher performance.

With second-generation Multi-Instance GPU (MIG), built-in NVIDIA Confidential Computing, and NVIDIA NVLink Switch System, H100 securely accelerates all workloads for every data center, from enterprise to exascale.

Supercharge Large Language Model Inference With H100 NVL

For LLMs up to 70 billion parameters (Llama 2 70B), the PCIe-based NVIDIA H100 NVL with NVLink bridge utilizes Transformer Engine, NVLink, and 188GB HBM3 memory to provide optimum performance and easy scaling across any data center, bringing LLMs to the mainstream. Servers equipped with H100 NVL GPUs
increase Llama 2 70B model performance up to 5X over NVIDIA A100 systems while maintaining low latency in power-constrained data center environments.

Enterprise-Ready: AI Software Streamlines Development and Deployment

NVIDIA H100 NVL comes with a five-year NVIDIA AI Enterprise subscription and simplifies the way you build an enterprise AI-ready platform. H100 accelerates AI development and deployment for production-ready generative AI solutions, including computer vision, speech AI, retrieval augmented generation (RAG), and more. NVIDIA AI Enterprise includes NVIDIA NIM™—a set of easy-to-use microservices designed to speed up enterprise generative AI deployment. Together, deployments have enterprise-grade security, manageability, stability, and support. This results in performance-optimized AI solutions that deliver faster business value and actionable insights.

Data sheet(s)