Praktikum: Evaluierung moderner HPC-Architekturen und -Beschleuniger (IN0012, IN2106, IN4294)

Dieses Praktikum wird nur in englischer Sprache angeboten.

Experimental Evaluation of modern Computing Systems and Accelerators 

This lab course will be held in English.

Bachelor/Master practical course (IN0012, IN4294 IN0014, IN2107, IN4602).

 

Organization

This lab is organized in collaboration with LRZ and LMU.

It is provided seperately for bachelor and master students, but carried out as one event.

Content

The LRZ runs an internal testbed environment with novel architectures and accelerators called BEAST (Bavarian Energy, Architecture and Software Testbed). This will help in shaping future large LRZ systems.

In this lab course, students will have the chance to work on newest HPC technology (such as CPUs and GPUs from Intel, AMD, NVidia, ARM), as they are available in the LRZ BEAST environment. Using small example workloads, the tasks in the lab will explore various features of novel technology on the node level. As tasks introduce new technology, the lab organizers will invite vendors to give talks on their newest hardware. The tasks will make use of different parallel programming models and constructs on the node level for multicore CPUs and accelerators, such as OpenMP (also for GPU), SIMD, synchronization. For later assignments, students are expected to make use of SYCL or CUDA.

The lab is split in two types of assignments: for the first part of assignments, students are expected to work on all architectures. For the second part of assignments, students work on specific architectures exclusively (randomly assigned to student groups).

Students will work in groups of 3 or 2, doing reports for each task as well as short presentations of their evaluation results, with 2 talks every week randomly selected among the groups. We use git (Gitlab) for code / report submission. For presentations, no slides are expected, but the report should be presented (MarkDown as shown by Gitlab WebUI).

Grading

Grading is done based on delivered reports and the short presentations. This includes source code used for the evaluations, which has to be added to the reports.

Registration

Registration for the lab will happen via the matching system.

Expected Skills

• Good knowledge of C on Linux
• Understanding of terms in computer architecture: CPU pipelining, SIMD, Out-Of-Order execution, multi-core, SMT, TLB, processor caches, replacement policies (before registering in the matching system, ask yourself if you can explain these terms!)
• Interest in computer architecture, benchmarking, and low-level code optimization

Schedule / Dates

Weekly meetings during the semester: every Thursday 4pm - 6pm, in FMI 01.06.020 (we may do some meetings at LMU, but not on first date)

First date: 24.4.2025