GPU Acceleration of HPC Applications with ISO C++ and CUDA C++

Content

In this course, you will learn how to accelerate HPC applications with CPUs and GPUs using the parallelism and concurrency features of the C++17, C++20 and C++23 standards as well as modern CUDA C++. We will demonstrate how to parallelize classic HPC patterns like multi-dimensional loops and reductions, and how to solve common problems like overlapping MPI communication with GPU computation. The material is supplemented with numerous hands-on exercises and illustrative HPC mini-applications. Along the way, you will learn about C++ concurrency features like threads, atomics, barriers, and parallel algorithms and how to integrate these into hybrid HPC applications using MPI.

Target Audience

This course is designed for CSCS users. Participants should have basic experience with C++11, C++ lambdas, and MPI.

Schedule

The course will begin on Tuesday, November 4, 2025, at 10:00 and conclude on Wednesday, November 5, 2025, at 17:00 (Central European Time, CET). Each day will include a one-hour lunch break, as well as a 15-minute break in both the morning and afternoon sessions.

Day 1: from 10:00 to 17:00

Intro and Modern C++ refresher
C++ standard parallelism and parallel algorithms
Data management
Ranges & views
Distributed Programming
Concurrency primitives
Execution Policies
Asynchronous execution
Case studies

Day 2: from 9:00 to - 17:00

Parallel algorithms
Asynchrony and streams
Cuda Kernels
Atomics
Synchronization
Shared memory
Cooperative algorithms
Grace-Hopper Architecture overview

Instructor

Dr. Benedikt Dorschner (Sr. Developer Technology Engineer, NVIDIA)

Participation Fee and Registration

All participants must register for the course. The registration fee includes lunch and coffee breaks.

Course Fee: CHF 160.00

Deadline for registration: Tuesday, October 28, 2025.

Kindly note that the course can take place only if there are sufficient confirmed registrations received by the deadline. The minimum number of participants is eight. Should the minimum number of participants not be reached by the deadline, registered attendees will be informed about the course cancellation. No communication will be sent to confirm the event. Registration for the course will automatically close when we reach the maximum number of participants (30).

Directions

Kindly note that no parking space is available at the Swiss National Supercomputing Centre. There is a Park & Ride Resega parking lot, within five minutes walk from CSCS.

You are encouraged to travel by public transportation. The closest bus stop to CSCS is "Lugano, Stadio." From Lugano railway station, you should take bus number 4 or 6.

Inquiries

Please contact Guilherme Peretti-Pezzi (guilherme.peretti-pezzi@cscs.ch) for questions related to the course content and info@cscs.ch for questions related to the event logistics.