OCF Training

This course introduces the fundamental concepts and tools required to write and run parallel code in Python. It is designed for participants who are already familiar with basic Python programming and want to leverage parallelism to improve performance in computational tasks.

Contents
The course covers the following key topics:

• Introduction to parallelism and concurrency
• Understanding the Global Interpreter Lock (GIL)
• Thread-based parallelism with threading
• Process-based parallelism with multiprocessing
• Parallel programming using OpenMPI and mpi4py
• Parallel data processing with joblib and dask.

Format
The course is hands-on and example-driven, with each concept introduced through live coding demonstrations and followed by guided practical exercises. Participants will write small parallel programs and observe performance differences in real time. The course assumes prior experience with Python basics, but no prior exposure to parallel computing is required.

• Prerequisites
  OCFP1

OCF Training

This course provides a practical introduction to Linux in the context of High-Performance Computing (HPC). It is designed for students, researchers, and professionals who are new to Linux. The focus is on command-line tools, scripting, system navigation, and the fundamental skills necessary to interact with HPC clusters.

Prerequisites
None

Contents
The course is structured to build proficiency in Linux, starting with the basics and progressing toward
more advanced HPC-related topics. Key areas covered include:

• What is Linux, and why are we using it?
• Setting up a Linux environment
• Basic Linux commands, file system navigation, file operations etc.
• Introduction to nano, editing files, saving files, creating files
• Basic shell scripts, setting environment variables, downloading and extracting files from the internet.
  Format

The course is delivered through a combination of a presentation and guided Linux tasks. Each module includes brief overviews followed by demonstrations and tasks. Exercises are designed to reinforce concepts and prepare participants for HPC usage.

Duration
This course is designed to be completed in approximately half a day (around 3 to 4 hours).

OCF Training

This course introduces the core concepts of a HPC cluster, and the basics of how to use it to submit jobs. It is aimed at users who are new to HPC (PhD students or industrial engineers) and need a foundational understanding of how to use a HPC system to conduct science/engineering.

Contents
The course covers the following key topics:

• Introduction to HPC, what it is, general concepts etc
• Accessing your/a HPC, ssh, ssh keys, graphical forwarding
• Copying data to and from the HPC and introducing the file system (what to save and where)
• Introduction to SLURM, queues, submitting a test job.
• Good practice on an HPC system.

Format
The course is delivered through a combination of a presentation and guided Linux/HPC tasks. Each module includes brief overviews followed by demonstrations and guided tasks. The end goal is to get users able to submit jobs to their HPC cluster.

Prerequisites
• OCFHPC1

OCF

This course provides a practical introduction to profiling and performance analysis of Python code. It is designed for Python users who want to identify bottlenecks, optimise runtime, and better understand how their code behaves under different workloads. Participants will learn how to use profiling tools to gain insights into function call frequency, execution time, and memory usage.

Contents
The course includes the following topics:

• The need for profiling: performance vs. readability trade-offs
• Built-in profiling tools: time, timeit, and cProfile
• Line-by-line profiling with line profiler
• Memory profiling with memory profiler and tracemalloc
• Best practices for writing performance-aware Python code

Format
The course is structured around live coding demonstrations and guided hands-on exercises. Each topic includes simple, real-world code examples designed to illustrate key profiling concepts.

Duration
This course is designed to be completed in approximately half a day (around 3 to 4 hours).

Prerequisites
• OCFP2

NVIDIA HPC, Software

Accelerating AI. Maximizing GPU Power. Unlocking Innovation

In the race to scale artificial intelligence, one resource stands at the centre: GPUs. Yet, too often, GPU power is locked away in silos, underutilized, or difficult to manage across teams. Run:AI changes that.

With Run:AI’s orchestration platform, enterprises gain a powerful layer of abstraction that makes GPU infrastructure work as flexibly and efficiently as the cloud. By pooling GPU resources and allocating them dynamically based on workload demand, Run:AI ensures your teams get the right compute, at the right time - with no wasted capacity.

Built for data scientists, researchers, and IT leaders, Run:AI simplifies the way organizations run AI workloads - whether training massive models, running inference at scale, or balancing multiple projects across teams.

Key Benefits:

• Maximize GPU utilization – Eliminate idle GPUs and reduce infrastructure costs
• Frictionless scaling – Run workloads across clusters, on-prem, cloud, or hybrid
• Team empowerment – Give every AI team on-demand access to the resources they need
• Simplified management – Centralize scheduling, monitoring, and policy control
• Future-ready infrastructure – Scale seamlessly as AI models grow in size and complexity.

From research labs to global enterprises, Run:AI unleashes the full potential of GPU infrastructure - accelerating innovation while reducing operational friction.

Run:AI: Orchestrate. Optimize. Accelerate.

Data sheet(s)

NVIDIA

Unlocking Real-Time Trillion-Parameter Models

GB200 NVL72 connects 36 Grace CPUs and 72 Blackwell GPUs in a rack-scale, liquid-cooled design. It boasts a 72-GPU NVLink domain that acts as a single, massive GPU and delivers 30X faster real-time trillion-parameter large language model (LLM) inference.

The GB200 Grace Blackwell Superchip is a key component of the NVIDIA GB200 NVL72, connecting two high-performance NVIDIA Blackwell Tensor Core GPUs and an NVIDIA Grace™ CPU using the NVIDIA NVLink™-C2C interconnect to the two Blackwell GPUs.

Data sheet(s)

Lenovo Components, HPC, Infrastructure, Network, Professional Services, Servers, Storage

OCF will leverage our deep expertise to design, deliver, deploy and support your next (or first) turnkey High Performance Computing (HPC) cluster based on Lenovo technology.

Every bespoke OCF / Lenovo cluster is built on proven technologies designed to maximise performance and optimise datacentre efficiency, tailored to your workloads and budget. Our solution design process is driven by a dedicated team of Presales Technical Architects who specialise in creating tailored HPC and storage architectures.

This team works in close collaboration with customers, technical partners and our in-house engineering teams to translate requirements into optimised, future-ready solutions. In addition to a broad market view of the industry and relevant technologies, the team is fully accredited and extremely experienced in architecting solutions based on Lenovo technologies.

For more information, read the PDF datasheet linked below or the CASE STUDY about Nottingham University.

Data sheet(s)

NVIDIA Infrastructure

AI has moved beyond training.

Today’s workloads demand real-time understanding, continuous reasoning, and multi-modal agentic behaviour - at massive scale.

Being the world’s most advanced rack-scale AI inference platform - purpose-built for the new generation of agentic AI, reasoning engines, and real-time multimodal intelligence.

As large language models grow from billions to trillions of tokens, training is no longer the bottleneck - reasoning is.

This is why NVIDIA built the GB300 NVL72: the world’s first AI inference platform engineered for real-time, high-context, multi-agent workloads.

Use Cases That Demand GB300 NVL72

• Real-time LLM inference with extreme throughput
• Multi-agent, multi-modal AI pipelines
• Next-gen digital twins and simulation AIs
• Autonomous control systems & robotics
• High-density, low-latency AI factories

Key Differentiators:

• Capability Traditional HGX GB300 NVL72
• Unified Memory ✘ ✔ (40 TB total)
• Power Smoothing ✘ ✔ (up to 30% reduction)
• Blackout Tolerance ✘ ✔ (BBUs built-in)
• AI Reasoning Optimized ✘ ✔
• Liquid Cooling Optional Standard