Is this event for me
This is the first iteration of a Machine Learning on GPUs course for the University of Manchester, University of Lancaster, University of Sheffield High Energy Physics, Nuclear Physics and Astronomy PhD students organised in collaboration between the 4IR Centre for Doctoral Training, University of Manchester Physics and Astronomy Department and the Training Working Group of the HEP Software Foundation.
- Get students comfortable and able to run ML jobs on GPUs
- Identify when to use GPU programming for tasks (which tasks are appropriate)
Some python programming background is expected.
Organisation and content
Due to the current pandemic circumstances we expect the whole event (course and training) to be held virtual. The link to the conference tool will be provided on this page.
The event will have 3 events spread over a week. There will be a kickoff event where the course and how to use the virtual material and get access to the GPUs will be explained (links are provided in agenda). Then 3 days for you to work on the material at your own pace. Finally 2 hands on sessions with tutors to go through the material interactively again.
Homework Study the material provided on Monday
|Homework Study the material provided on Monday
||Homework Study the material provided on Monday
||ML hands on
break out sessions
GPU hands on
It is important that you do your homework to make the hands on sessions more productive.
The number of participant places for this course and training is limited to 36 i.e. the number of GPUs available.
In case you registered and got a place but find out later that you cannot attend, please contact the organizers so we eventually could make your place available to others.
The content will be available online for those who cannot participate and in the future there might be further hands-on training.
- Anna Scaife (University of Manchester)
- Meirin Oan Evans (University of Sussex)
- Savannah Thais (Princeton)
- Stephen Menary (University of Manchester)
Cost of Attendance
The training is organized free of charge. The only thing we ask is that if you register to attend, you commit to attending the full event.