TDDC78 mid-term evaluation 2013The course has been mid-term-evaluated by the muddy card method on wednesday 24/4/2013 during the 9th lecture.
32 students attended the lecture, and I received 31 cards.
Subject, general issues
- Interesting / relevant subject (6)
- Would like to know what would come in the exam (1)
- Excursion to NSC was appreciated (7)
- Fun to use actual [supercomputer] hardware (2)
- pthreads is not covered but not everybody has worked with it before (1)
- Lab sessions should have started earlier, even without using Triolith (1)
- Could test first locally at IDA (2)
- Good lectures (3)
- Interesting lectures (4)
- Appreciate the professor's dedication (1)
- Very good explanations (1)
- Well structured lecture material (1) / Very good slides (1) / Good slides, a bit theoretical (1)
- Cover a lot (2)
- A lot of theory, too few examples (1)
- Sometimes a bit too cluttered slides (1)
- Too many slides per lecture (1)
- Sometimes run out of time (1)
- First lectures should be expanded and take up [even] more on design, architecture, communication etc. (1)
- Lectures and lessons could take up more on implementation and problem solving (1)
- Could give more concrete examples of using OpenMP, MPI, tools (1)
- Examples are a bit too simplified (1)
- Could give more implementation / code examples, esp. OpenMP (3)
- Sometimes a bit slow (1)
- Talk slower (1)
- Sometimes speaking to the wall (1)
- More examples on the whiteboard (2)
- Very interesting NSC introduction (1)
- very useful (1)
- good with code examples and lab tutorials (1)
- are basically code lectures, good complement of the lectures that do not go through so much code (1)
- useless as they were mostly duplication of the lecture contents (3)
- Would like to solve small problems, more discussions, not another lectures. (2)
- Lesson slides are useful but should be available earlier. (1)
- Good, useful, interesting, fun (16)
- Useful to understand the theory (3)
- Challenging (1)
- Realistic problems to solve (3)
- Good lab compendium (1)
- Lab compendium should be more detailed (1) / lab instructions were insufficient (1)
- Good to start from a given sequential code (2)
- Threshold for the labs is set very high; difficult in the beginning (2)
- Would be good with an intro lab [Hello World on Triolith] (3)
- Lab 1 should be broken down into smaller and easier tasks (1)
- Lab 1 should give the solution of the problem so we can focus on MPI (1)
- The difficulty in Lab1 is not in MPI itself but in using C. (1)
- Problems with makefiles, linking, modules etc. (1)
- Too few lab assistants (2)
By and large, the course seems to run very well.
Whenever I talk too fast please feel free to interrupt and let me know.
Note that the slide sets do not generally coincide with lecture slots. It is normal that some slide sets (e.g. MPI, OpenMP) span more than 90 minutes.
The lecture material is gradually being revised as far as my time permits. We will try to include more code examples where appropriate, but there is certainly a limit to how much code can reasonably be taken up in the lectures. We can (unfortunately) not increase the number of lecture slots significantly because the schedule frame, the course credit point volume and the course budget set the constraints.
The suggestion of having an introductory Lab 0 (Hello World on Triolith) raised by several students looks very interesting; we will discuss this for next year's instance of the course. We will also consider a revised setup for the lessons next year, which is in line with some of the suggestions given above.
Thanks for all comments!
Christoph Kessler, course leader TDDC78
Page responsible: Webmaster
Last updated: 2013-04-30