Skip to main content

Ctrl+K

GitHub Discussion
Pages

About

About this Seminar

Part 1: Utilizing Shared Resources

1. What challenges did you run into?
2. Challenges in Computational Projects
3. Primer on Computer
4. From Megahertz to Teraflops
5. Concurrency
6. From Concurrency to Service
7. as a Service
8. Virtualization
9. The Cloud
10. HPC Clusters
11. Distributed Storage

Part 2: Research Software Engineering

1. Computational Project
2. Introduction to Version Control
3. Project Documentation
4. Extended Documentation
5. Managing Data
6. Code Structure
7. Environments 1.0
8. ✏️ Project Exercise 1 ✏️
9. Primer on Software Development
10. ✏️ Project Exercise 2 ✏️
11. Software Development Strategies
12. Refactoring
- 12.1. Definition
13. Notes on Python
14. Primer On Parallelism
- 14.1. Primer on Parallelism
- 14.2. Parallelism in Python
15. ✏️ Project Exercise 3 ✏️
16. Environments 2.0
17. 📝 Project Homework 📝

Part 3: Recap; Containers, Clouds & Objects

1. Recap: Last Week
2. Recap: exoHelloWorld
3. Apptainer Runtime Guide
- 3.1. run vs exec vs shell
4. Suggested container definition
5. Containers for SoC projects
- 5.1. Environment Management
- 5.2. Concern Substitution
6. CI/CD Build
7. Fetching a Container at Runtime
8. The UZH Science Cloud
9. Ansible Detour
- 9.1. Getting Started with Ansible
- 9.2. One Shot Config
10. Swift Object Storage

Part 4: Cluster - Workflow & Profiling

1. Recap: HPC Clusters
2. Containers for HPC
3. HPC Clusters at UZH
4. Login and Look Around
- 4.1. Helpful Commands
5. ✏️ Exercise 1 ✏️
6. Latency and Throughput
- 6.1. Connection Fabric
- 6.2. GPU-NVMe Latency Comparison
7. Profiling
5. ✏️ Exercise 1 ✏️
8. 📝 Project Homework 📝
9. Q & A
10. Special Topics
- 10.1. Using Shared Memory (/dev/shm)

Outro

Outro

Repository
Suggest edit

.md

The Industry Replied

4.2. The Industry Replied#

Power Wall

Instead of making a single core faster, simply add more cores!

previous

4.1. How kept computational power growing?

next

4.3. Challenges

By Jonas Liechti - https://github.com/j-i-l, Matteo Delucchi - https://github.com/matteodelucchi

© Copyright 2026, T4D GmbH.