Q13 · Year 2 — Advanced Engineering & Specialisation · Year 2, weeks 23–28
Distributed HPC and Hybrid Orchestration
Year 2, weeks 23–28: HPC architecture and cluster engineering; MPI and distributed scientific computing; GPU kernel engineering; Distributed quantum simulation; Tensor network algorithms; Hybrid quantum-classical orchestration.
Topics
Scientific concepts
- 01NUMA
- 02interconnects
- 03scheduling
- 04Collectives
- 05domain decomposition
- 06scaling studies
- 07Memory coalescing
- 08occupancy
- 09mixed precision
- 10Partitioning
- 11communication overhead
- 12scaling
- 13Contraction ordering
- 14Workflow engine
- 15classical feedback
- 16experiment pipelines
Computational methods
- 01distributed profiling
- 02profiling
- 03tensor networks
- 04MPS/PEPS concepts
- 05benchmark suite
Practical implementation
- 01Derive and model — derivation and problem set
- 02Software build — a tested, documented, version-controlled code module
- 03AI / HPC / algorithms — benchmark against baselines for accuracy, runtime and memory, with limitations stated
- 04Engineering systems — map the topic to hardware, materials, control or architecture constraints (design note or block diagram)
- 05Lab / simulation — a simulation, remote quantum-hardware run or supervised practical, with data and uncertainty captured
- 06Research / integration — read or reproduce one relevant result and integrate it with the team workstream (logbook and review)
Project
Module portfolio contributing to the Year 2 outputs: Specialisation portfolio; compiler/runtime or device/control prototype; advanced QEC decoder; experimental or HPC benchmark; interface-controlled subsystem.
Expected competencies
- 01Derive, implement, benchmark and experimentally test hPC architecture and cluster engineering
- 02Derive, implement, benchmark and experimentally test mPI and distributed scientific computing
- 03Derive, implement, benchmark and experimentally test gPU kernel engineering
- 04Derive, implement, benchmark and experimentally test distributed quantum simulation
- 05Derive, implement, benchmark and experimentally test tensor network algorithms
- 06Derive, implement, benchmark and experimentally test hybrid quantum-classical orchestration
Assessment
- 01Weekly review — research log, code repository, simulation notebook, experiment records, design artefact and oral explanation
- 02Six-week milestone review — mathematics, software quality, computational performance, experimental integrity, hardware awareness, documentation and research contribution
- 03Module portfolio — the outputs of every week in the module
Source document: FQSE Three Year Master Timetable TUTOR Maroon