Introduction to GPU programming: When and how to use GPU-acceleration? 21 April 2020
GPU-acceleration has been shown to provide significant performance benefits in many different applications. However, for a novice, or even for a moderately experienced scientist or programmer, it is not always clear which applications could potentially benefit from GPU-acceleration and which do not. For example, a Nvidia V100 GPU can perform artificial intelligence (AI) related computations in a fraction of the time it takes a regular CPU to perform the same computations but ill-informed OpenACC compiler pragma can actually make a code run slower. Why is this? When should one invest time in GPU-acceleration? How much speedup can be expected with a given application?
The main goal of this one day course is to start answering these questions. The course also covers the basics of GPU programming and aims to provide the necessary information for avoiding the most common pitfalls.
- The course does not require any existing GPU programming knowledge but basic understanding of the C language is required for the hands-ons
- For the hand-on practicals participants should bring a laptop. The laptop should be able to access the internet via eduroam (needs to be arranged in your home institution) and have a ssh-client installed
- To be able to login to Vera and do the exercises, all participants should register in the SUPR-portal and take the necessary steps to create an account. More information will be sent out to the participants before the course starts.
The teaching language will be English.
The course forms part of the SNIC coordinated training in HPC. It is open to users from all six SNIC HPC centres and academics throughout the Nordic region.
Time and location
NOTE Due to the current situation with the COVID19, the course is postponed to the autumn. Registration is currently not possible. Updated information will be announced later.
The course will be held at the Johanneberg campus of Chalmers University of Technology in the room "PJ" (building Fysik Origo, entrance from Fysikgården 2B, Entrance floor).
NOTE: registration is binding. Should you not be able to participate you must inform C3SE by sending and email to firstname.lastname@example.org no later than 7 days before the course starts. Charges may apply otherwise.
To register, please fill in the registration form:
Deadline for registration: 2020-04-15
Questions and queries regarding the course should be directed to C3SE-support.