Many thanks to all of you who took part in the course, and especially many thanks to all of you who gave some course feedback! Overall, 102 students passed the course and 98 students gave some course feedback — see below for more details. I am very pleased that so many of you gave course feedback, and especially that many of you also took some time answering the free-form questions.
This year, we had revised many aspects of the course based on the feedback that we got in the 2015 edition of this course. In particular, we had revised the overall schedule of the course, improved the course material (especially parts related to GPU programming), and thoroughly redesigned the grading and submission system. Based on the feedback that we got, these were the right decisions, and the overall structure and the workflow of the course are in a fairly good shape now. However, we have also idenified many details that we can try to improve for the next year.
Thanks, and see you again!
— Jukka Suomela, on behalf of the whole course team.
See the exercise statistics page for more information on the overall progress during the course, as well as for statistics related to individual exercises.
We got course feedback as follows:
The course feedback seems to represent well all students who took part in the course. More than 91% of all students who passed the course gave some feedback, we got plenty of feedback from students with both high grades and low grades, and we also got some feedback from several students who did not pass the course.
|≥ 1 point||≥ 30 points||≥ 50 points||≥ 60 points|
|physics & math||12||11||8||1|
(Only those categories shown with at least 4 students with the grade of 5/5.)
Overall, I am very happy to see that most of the students found the course interesting and well-implemented. Many of the improvements that we implemented this year were clearly successful.
Some of the areas that need improvement are related to aspects that we are still developing. For example, we thoroughly revised CUDA-related material for this year, and based on the feedback we are going in the right direction, but some further work is still needed. Also the need for providing more help related to debugging and profiling tools was already identified last year, and while we provided this year more information on debugging tools, clearly we can try to do better next year.
There were lots of comments on the schedule and workload. I would like to point out that the schedule is a bit tight by design: our goal is to have an intensive course that entirely fits in the 5th period. In general, most CS courses in Aalto seem to be organised in periods 1–2 or 3–4, and students have been asking for something to study for the 5th period. This course tries to fill in this gap, but this also means that all course activities are packed in only 6 weeks. However, the uneven workload is something that we can try to fix for the next year. In particular, we are planning to develop additional optional exercises that can be solved in a flexible manner during those weeks when the workload is lower.
Maari computers are currently one of our main technical challenges. We are very well aware of the issue. Unfortunately, it is not easy to find a cost-efficient way of organising a dedicated platform for benchmarking code for one hundred students. We will nevertheless work together with Aalto IT and also with CSC in order to find a better solution.
I think one of the main areas that we will need to focus on next year is providing more support for beginners. We can also try to develop a preprequisite test that is better aligned with the requirement level of the course.