|Date:||30.10.2018 13:00 - 30.10.2018 14:00|
|Location details:||Join the webinar using Zoom. Link: https://cscfi.zoom.us/j/460246394|
|Lecturers:|| Pekka Manninen (CSC) |
The reason for doing high-performance computing is, well, high performance - but is your application as good as it can get, in terms of performance? Watch this webinar series and you will have an answer to that - hopefully together with some ideas on how to improve the observed shortcomings.
In the first part, we will discuss some general considerations related to application performance, followed by a walkthrough on how to profile and analyse performance of your parallel application on CSC's Cray XC supercomputer Sisu. The second part will cover the foundations as well as programming practices for writing efficient program code for modern x86 CPUs. In the last part, we will talk about some algorithmic considerations and MPI programming practices for improving parallel scalability of MPI-parallel applications.
30.10. Performance optimization of scientific software I: Performance analysis
6.11. Performance optimization of scientific software II: Node-level performance tuning
20.11. Performance optimization of scientific software III: Improving application scaling
Join the webinar with Zoom: https://cscfi.zoom.us/j/460246394
Tips: At the first time, Zoom asks you to download a launcher plug in. After this, you should be able to open the webinar room from the link. The webinar room is timed, and will open when the webinar begins. When signing in, remember to click "join audio conference by computer". If you are using a headset, it is easiest to plug them in before opening Zoom.