The instructor is explaining in a voice and manner that puts you to sleep easily.

The course materials are not well-prepared. As a student without prior knowledge of GPU programming, I find it very hard to follow. Also, the instructions of assignment are confusing and unclear, I even couldn't know what I should do. I tried to find tips from discussion forum, but I realized that lots of students' questions there didn't receive any response from TA or staff .

Overall, this course makes me frustrated and isn't worth to me to continue my subscription anymore.

Assignments really need to be worked on. There are so many errors and the overall objective of the assignments in not clear. Instead of working on the concepts, I had to work on how to run files in the assignment.

The theory and explanations provided is also not adequate.

Lectures are terrible: no real explanation of what is going on, just definition of terms. Some links to reference material are broken. Presenter's voice is not very clear: speaks in low vocal fry (i.e. mostly whispered). Example code in labs have minimal explanation. The examples are not explained in lectures. This is stuff we'd expect just from a brief Google search.

More about concepts of threading, but not about gpu\cuda programming. It's ok, but if you are familiar with these concepts you could skip this course

Labs sometimes have some little bugs, but could be resolved easily based on hints from corresponding discussion section

Very slow pace but otherwise a great learning opportunity, especially the lab resources

Very helpful. Learned something basic that was very important. Many thanks!

Clearly provided material with labs and quizzes that are well organized.

Great introductory course, well paced.

Zoʻr

GOOD

Overal a good class, very useful introduction to CUDA including lots of key information - It would be very beneficial to have some more information and examples on the parallel classes, mutexes, atomics, asyncio etc. - Some of the existing, provided code in the assignments can be a bit confusing because it seems as if the assignment is expecting more than is required. There were a few comments about this in the forums I think

I learnt a lot, and exactly the stuff I wanted to learn + some things that were unknown unknowns to me.

Would give 5 stars, but I think it's missing a bit of a polish, some exercises are not as clearly stated as they could be, and there are links to additional resources that are still broken. I think it is still worth doing the course despite all this, but I would like to see the course administration improve these things.

Super low energy presentation, and the material seems largely cut and paste from NVIDIA and other online sources, which I already can find and organize quickly using ChatGPT. This first course was "ok", but didn't really add much momentum to my learning.

The teaching is good, but there are quite a few mistakes in the task instructions, especially in the last ones. Please remember that we are students, and any mistake you make can confuse us and slow down our learning.

good course to boost your skills

It was mostly about how to compile and link CPU/GPU files. I was expecting that they will teach how to write some parallel programs on GPUs.

Not a good course. The content are in a poor quality and also the instructor does not take much effort to clearly explain things.

not properly guided submissions become a bit difficult

the assignment is too obscure

Most of the information covered could be easily found in documentation and was a bit too surface level for my taste. To truly explore the concepts presented in the course, I needed to work through my own examples rather and I found the assignments unhelpful in learning the material. The assignments tended towards "fix these five things in the code" rather than having me grapple the presented information which I think is necessary when learning . The information was clearly presented, I just found the course to be far too easy.

Honestly, not the quality I was expecting from JHU. Courses I've done from JHU before are significantly better. The assignments we're basically half effort cobbled together. Ironically, the instructor talks about having a good README in the project structure, the README files in the assignments are just copied over from assignment to assignment, and thus don't even make sense.

The coding assignments are poorly developed. The automated grader fails with no helpful message on a few assignments, to the point that you need to reverse engineer the thing to find exactly what needs to be submitted. The problem description simply does not tell what the actual answer should be. It was a frustrating experience.

The quality of this course is quite mediocre, with many errors. Moreover, the cited sources are nothing more than Wikipedia and GeeksforGeeks, which lack academic authority. Not to mention, the teacher's PPTs are very poorly made, and the quizzes are largely unrelated to the lecture content.