Apply now
Master'sData, AI, Simulation & Analytics Applications open

Master’s in HPC & GPU Engineering for Simulation and AI

Online master's programme of 1,900 hours in HPC and GPU engineering for simulation and AI. 12 modules covering GPU architecture, CUDA programming, performance portability, distributed computing, GPU-accelerated simulation, distributed AI training, inference optimisation, cluster operations and energy efficiency under ISO/IEC 30134-2, ISO/IEC 9899 and ISO/IEC 14882, with an assigned technical tutor.

Duration1900 h
ModeOnline
LanguageEnglish
Start datesSet by the faculty
AwardSEIUM Institute diploma
Enrol now View syllabus Ask an advisor
Virtual lab · try it before you enrol

Take the controls.

Run a real engineering trade-off from this field. Move the sliders, watch the model react in real time and beat the challenge.

SIM · LIVE ●
Learning outcomes

What you will be able to do

01

Understand GPU architectures and HPC systems.

02

Programme efficient compute kernels with CUDA.

03

Port applications between acceleration platforms.

04

Scale simulations and training runs to multiple nodes.

05

Optimise AI model inference on GPUs.

06

Operate clusters with resource managers and containers.

07

Assess the performance and energy efficiency of computing.

Profile

Who it's for

  • HPC engineers moving their workloads onto GPU-accelerated systems.
  • CUDA developers who want to master multi-GPU and multi-node scaling.
  • AI infrastructure engineers running large training and inference platforms.
  • Performance engineers optimising scientific and AI codes.
  • HPC system administrators operating GPU clusters.
  • Computational scientists accelerating their simulations.
  • Engineering or science graduates looking to enter research.
Standards & tools

What you'll work with

  • ISO/IEC 30134-2 (data centre power usage effectiveness, PUE)
  • ISO/IEC 9899 (C programming language)
  • ISO/IEC 14882 (C++ programming language)
  • ISO/IEC 30134-1 (data centre key performance indicators)
  • UNE-EN 50600-1 (data centre facilities and infrastructures)
  • ISO 50001 (energy management systems)
  • Directive (EU) 2023/1791 (Energy Efficiency Directive)
  • CUDA Toolkit
  • NVIDIA Nsight
  • PyTorch
  • Slurm
  • Kokkos
Syllabus

12 modules · 1900 h

  1. 1.1Evolution of supercomputing
  2. 1.2The TOP500 and Green500 lists
  3. 1.3Exascale computing and its challenges
  4. 1.4Convergence of HPC and AI
  5. 1.5The European EuroHPC infrastructure
  6. 1.6Access models: on-premise and cloud
  7. 1.7Simulation workloads versus training
  8. 1.8Total cost and energy efficiency
  9. 1.9Hardware trends
  10. 1.10Case study: sizing an R&D cluster
  1. 2.1Multiprocessors, warps and threads
  2. 2.2Memory hierarchy: global, shared and registers
  3. 2.3Tensor Cores and mixed precision
  4. 2.4HBM bandwidth
  5. 2.5NVIDIA and AMD architectures
  6. 2.6NVLink and PCIe interconnects
  7. 2.7Occupancy and latency
  8. 2.8The roofline model for GPUs
  9. 2.9GPU versus CPU by workload type
  10. 2.10Lab: measuring bandwidth and peak performance
  1. 3.1Kernel programming model
  2. 3.2Memory management and transfers
  3. 3.3Shared memory and coalescing
  4. 3.4Synchronisation and atomic operations
  5. 3.5Streams and concurrency
  6. 3.6Parallel reductions and prefix sums
  7. 3.7cuBLAS, cuFFT and cuSPARSE libraries
  8. 3.8Debugging with Compute Sanitizer
  9. 3.9Profiling with Nsight
  10. 3.10Practical: optimised matrix multiplication in CUDA
  1. 4.1HIP and ROCm
  2. 4.2SYCL and oneAPI
  3. 4.3GPU offloading with OpenMP
  4. 4.4Acceleration directives with OpenACC
  5. 4.5Kokkos and RAJA
  6. 4.6Accelerated Python with Numba and CuPy
  7. 4.7JAX and XLA compilation
  8. 4.8The Triton language for kernels
  9. 4.9Performance portability strategies
  10. 4.10Workshop: porting a CUDA kernel to SYCL
  1. 5.1MPI and point-to-point communication
  2. 5.2Collective communications and their cost
  3. 5.3MPI with GPUs and CUDA-aware MPI
  4. 5.4NCCL for inter-GPU communication
  5. 5.5InfiniBand networks and RDMA
  6. 5.6Network topologies
  7. 5.7Domain decomposition on GPUs
  8. 5.8Overlapping computation and communication
  9. 5.9Scaling to multiple nodes with MPI
  10. 5.10Lab: multi-GPU stencil with MPI
  1. 6.1Dense and sparse linear algebra
  2. 6.2Iterative solvers on GPUs
  3. 6.3GPU-accelerated CFD
  4. 6.4Lattice Boltzmann methods
  5. 6.5GPU-accelerated molecular dynamics
  6. 6.6Finite elements on GPUs
  7. 6.7SPH particle methods
  8. 6.8Mixed precision in simulation
  9. 6.9Validating accelerated results
  10. 6.10Practical: Lattice Boltzmann solver on a GPU
  1. 7.1Data parallelism with DDP
  2. 7.2Model and tensor parallelism
  3. 7.3Pipeline parallelism
  4. 7.4FSDP and ZeRO
  5. 7.5Mixed precision and BF16
  6. 7.6Activation checkpointing
  7. 7.7Optimising the data pipeline
  8. 7.8DeepSpeed and Megatron frameworks
  9. 7.9Scaling large models
  10. 7.10Lab: multi-GPU training with PyTorch FSDP
  1. 8.1GPU inference and latency
  2. 8.2TensorRT and graph compilation
  3. 8.3INT8 and FP8 quantisation
  4. 8.4Triton inference servers
  5. 8.5Dynamic request batching
  6. 8.6Inference of large language models
  7. 8.7KV cache and paged attention
  8. 8.8Optimising attention kernels
  9. 8.9Inference performance metrics
  10. 8.10Workshop: optimising a model with TensorRT
  1. 9.1Resource management with Slurm
  2. 9.2Kubernetes for GPU workloads
  3. 9.3Apptainer containers on clusters
  4. 9.4Scheduling mixed workloads
  5. 9.5GPU monitoring with DCGM
  6. 9.6Lustre and GPFS file systems
  7. 9.7Storage for AI
  8. 9.8Security in HPC environments
  9. 9.9User and quota management
  10. 9.10Practical: deploying GPU jobs with Slurm and Apptainer
  1. 10.1Profiling complete applications
  2. 10.2Memory and compute bottlenecks
  3. 10.3I/O optimisation
  4. 10.4GPU energy consumption
  5. 10.5Power and frequency capping
  6. 10.6Liquid cooling and PUE
  7. 10.7Carbon footprint of computing
  8. 10.8Data centre efficiency metrics
  9. 10.9HPL, HPCG and MLPerf benchmarks
  10. 10.10Lab: energy efficiency of a training run
  1. 11.1Accelerated surrogate models
  2. 11.2Neural operators for simulation
  3. 11.3Real-time digital twins
  4. 11.4Generative design with GPUs
  5. 11.5Climate and weather simulation
  6. 11.6Life sciences and computational chemistry
  7. 11.7Autonomous driving and sensor simulation
  8. 11.8Hybrid quantum computing
  9. 11.9Industrial acceleration cases
  10. 11.10Case study: neural surrogate of a CFD model
  1. 12.1Selecting the application
  2. 12.2Initial performance analysis
  3. 12.3Designing the parallelisation strategy
  4. 12.4GPU implementation
  5. 12.5Multi-node scaling of the application
  6. 12.6Optimisation and profiling
  7. 12.7Efficiency assessment
  8. 12.8Conclusions on the performance achieved
  9. 12.9Master's Thesis report
  10. 12.10Master's Thesis: defence before a panel
Methodology, software & labs

How you'll learn

100% online training on the SEIUM campus, with downloadable material, recorded classes and an assigned technical tutor. There are no fixed timetables: you progress at your own pace within the intake, so you can combine the programme with your job.

Each module combines theory, guided reading of the original standard and applied exercises. The working software is CUDA Toolkit, NVIDIA Nsight, PyTorch, Slurm and Kokkos, in free, academic or evaluation versions. Real procedures are studied: optimising compute kernels on GPUs, distributed training of AI models, and energy efficiency studies on a cluster. Assessment is continuous by module, plus the final Master's Thesis.

Standards covered: ISO/IEC 30134-2, ISO/IEC 9899, ISO/IEC 14882, ISO/IEC 30134-1, UNE-EN 50600-1, ISO 50001 and Directive (EU) 2023/1791 (energy efficiency).

Assigned technical tutor:every module ends with a deliverable that your tutor marks individually.
Standards-based learning:guided reading of ISO/IEC 30134-2, ISO/IEC 9899 and ISO/IEC 14882 applied to real cases, not as a theoretical annex.
Industry software:CUDA Toolkit, NVIDIA Nsight, PyTorch, Slurm and Kokkos, in free, academic or evaluation versions.
Master's Thesis with defence:the GPU acceleration and multi-node scaling of a real simulation or AI application, defended before a panel.
Flexible format:100% online, no fixed timetables, compatible with full-time work.
Capstone project

Build it end to end

Your final project follows the real engineering workflow, step by step, reviewed by your tutor.

Profile the application
Select the application and analyse its baseline performance.
Design the parallel strategy
Define how the workload will be parallelised across GPUs.
Implement on GPU
Port the compute-intensive parts to CUDA or a portable model.
Scale and optimise
Scale to multiple nodes, profile with Nsight and assess energy efficiency.
Master's Thesis
Present the performance achieved and conclusions, defended before a panel.
Career opportunities

Where this takes you

HPC EngineerCUDA DeveloperAI Infrastructure EngineerPerformance EngineerHPC Systems AdministratorComputational Scientist
Admissions & scholarships

Requirements

  • University degree in engineering, science or architecture, or equivalent demonstrable professional experience.
  • Internet connection and a computer capable of running the programme software.
  • English sufficient to read technical standards.

Ready to start?

Send your application today. Our admissions team replies within 72 hours with your study plan and funding options.

Apply now Scholarships
FAQ

Frequently asked questions

The programme totals 1,900 hours over 12–14 months. You set your own pace within each intake, and your tutor helps you plan the workload around your job.

You work with CUDA Toolkit, NVIDIA Nsight, PyTorch, Slurm and Kokkos, using free, academic or evaluation versions. Your tutor indicates the free alternative wherever one exists.

A SEIUM Institute diploma, awarded on passing the continuous assessment and the Master's Thesis. It is SEIUM's own qualification, not an official state degree; its recognition depends on the programme completed and the regulations of each country.

Yes. Modules passed in a course or postgraduate diploma in the same area are recognised in the higher programme. Ask about the recognition policy when you enrol.

A university degree in engineering, science or architecture (or equivalent professional experience), a computer able to run the programme software, and English sufficient to read technical standards.

Start dates are set by SEIUM's faculty for each intake. Our admissions team will confirm the next available date when you apply.

Keep exploring

Related programmes

See all in Data, AI, Simulation & Analytics
Master's7,730 $

Master’s in Synthetic Data and Sensor Simulation Engineering

Online master's programme of 1,900 hours in synthetic data and sensor simulation engineering. 12 modules covering computer graphics for simulation, camera, LiDAR and radar modelling, environment…

1900 hOnlineData, AI, Simulation & Analytics
Master's11,090 $

Master’s in Edge AI & TinyML Engineering

Online master's programme of 1,900 hours in edge AI and TinyML: 12 modules covering microcontrollers and embedded systems, efficient neural networks, model compression, frameworks and compilers,…

1900 hOnlineData, AI, Simulation & Analytics
Master's11,090 $

Master’s in Model Verification, Validation and UQ Engineering

Online master's programme of 1,900 hours in model verification, validation and uncertainty quantification (UQ) engineering. 12 modules covering model credibility, code and solution verification, uncertainty propagation,…

1900 hOnlineData, AI, Simulation & Analytics
Master’s in HPC & GPU Engineering for Simulation and AIEnrol
Related programmes

More in Data, AI, Simulation & Analytics

Master’s in Model Verification, Validation and UQ EngineeringView programme →Master’s in Optimisation and OR for EngineeringView programme →Master’s in Simulation, Co-Simulation and HPC EngineeringView programme →Master’s in Real-Time Analytics and Stream Processing EngineeringView programme →Master’s in Multiphysics Co-Simulation & FMI EngineeringView programme →Master’s in Federated Learning & Privacy-Preserving ML EngineeringView programme →

All Data, AI, Simulation & Analytics programmes → · How to apply · Fees & financing · Scholarships

Scroll to Top
SEIUM – Institute of Advanced Engineering
Powered by  GDPR Cookie Compliance
Privacy Overview

This website uses cookies so that we can provide you with the best user experience possible. Cookie information is stored in your browser and performs functions such as recognising you when you return to our website and helping our team to understand which sections of the website you find most interesting and useful.