CP2K-UK still going strong
Posted: 28 Sep 2016 | 15:06
September seems to have passed by in a bit of a blur, and it's already a whole month since the CP2K Summer School, which we ran at King's College London (23-26th August), so I thought it would be a good time to give an update on the recent activities of the CP2K-UK project.
One of the things I always hear when talking to people about CP2K is that they want more access to CP2K training events.
Every two years there has been a CECAM-sponsored CP2K Tutorial in Zurich, so we took the opportunity to run a similar event here in the UK. The 3-day summer school proved to be really popular, with 50 researchers (about half from the UK) attending. King's College London (one of the CP2K-UK project partners) provided the venue at the Strand campus in the centre of London, with a spacious lecture theatre and well-equipped computer lab next door. This made it really easy to intersperse lectures covering the basics of CP2K as well as some specific advanced functionality like TD-DFT and RPA with hands-on sessions in the lab.
We had a good mixture of lecturers - including myself, Matt Watkins and Lianheng Tong from the CP2K-UK project, Sanliang Ling from UCL and Jan Wilhelm from Prof. Jürg Hutter's group at University of Zurich. Sergey Chulkov from Lincoln also helped out during the lab sessions as we were kept busy with lots of questions and discussion. We had access to ARCHER too, so participants could try running CP2K at scale, and everything worked really smoothly.
The only downside of the week was that it coincided with a heat-wave, and the air conditioning in the lab struggled to cope at times! As well as the sessions during the day, thanks to sponsorship from Nanolayers and the CritCat project, we were also able to have a social dinner at the excellent Belgo restaurant near Covent Garden - with copious quantities of rotisserie chicken, moules frites and Belgian beer.
93% of respondants to our feedback survey said the school was "Excellent", "Very Good" or "Good" and most importantly, 76% agreed that they now knew enough about CP2K to use it in their research - so mission accomplished! We also got some nice comments including "a big thank you to all organisers and lecturers for all their work and assistance!".
Several people requested that we cover a wider range of topics and also cover more application-oriented talks, so when we run another school in future I think we might need to make it 5 days long to fit everything in. All of the material from the school is available on the CP2K website for everyone to make use of.
Aside from the summer school, we have several changes of staff on the project. Lianheng Tong, who has been based at King's College with Prof. Lev Kantorovich, is leaving academia to take his software development and training skills into SMEs. Sergey Chulkov joined the CP2K developer team in Dec 2015, working with Matt Watkins at Lincoln on the TD-DFT code (thanks to ARCHER eCSE funding) and has recently implemented support for hybrid functionals and the Maximum Overlap Method. It looks likely that he will be able to continue for another year to work on CP2K's device electronics modelling code, which is great news!
Here at EPCC we have just completed work on a tree-based CP2K input editor, and will say goodbye to Jeremy Nowell as he moves on after 14 years at EPCC. Gordon Gibb will be joining the project and working on integration of some CP2K support into the Chimera, as well as exploring CP2K performance on our new Intel Broadwell cluster, Cirrus. Mark Tucker is also continuing until at least the end of the year working on improving CP2K performance, with lots of new OpenMP being added to the code week by week.
We don't have any firm dates for further CP2K events, but will soon be planning a user group meeting in early 2017, so keep an eye out for that.
Iain Bethune, EPCC