GPU and Distributed Architecture Computing 2024-2025    1st Semester

Instructor:

Conf. Dr. Rareș Boian <rares at cs.ubbcluj.ro>

Syllabus

Online Communication and Collaboration

• Join the Microsoft Teams team with code 7lw1l4y
• For announcements watch:
  • Your university email @stud.ubbcluj.ro
  • Microsoft Teams General channel.
• The discussion forums will be in Microsoft Teams.

Graded Work

• Projects
  • GPU (G): GPU calculation of proof-of-work nonce
  • Threads/Actors (T): User-level threads and actors library
• Written exam (E)

Final grade calculation

• In order to take the writen exam or the re-exams and get a final grade, all projects must be presented
• Project delays result in a 1 point penalty per week, not including vacation weeks
• Grade calculation formula:
  
  
  Grade = 1 + 0.3*G + 0.3*T + 0.3*E

Re-examination

• Present the projects not yet presented
• Grades will take into account the delay penalty above
• To improve a project grade, discuss extra work with the instructor

Weekly Schedule

Week	Lecture	Seminar	Projects
			G	T
1: Sep 30	Parallel architectures; GPU project kick-off	Project development
2: Oct 7	GPU architecture; parallel algorithm examples
3: Oct 14	Introduction to CUDA	Project development
4: Oct 21	Introduction to OpenCL
5: Oct 28	GPU work scheduling	Project development and grading
6: Nov 4	GPU memory usage optimizations
7: Nov 11	Synchronization, atomic operations, streams	Project development and grading
8: Nov 18	GPU project (F, N) optimization analysis
9: Nov 25	Distributed programming paradigms overview	Project development and grading
10: Dec 2	Functional and Clojure paradigms
11: Dec 9	Communicating sequential processes	Project development and grading
12: Dec 16	Actors and agents
-: Dec 23	Vacation
-: Dec 30	Vacation
13: Jan 6	Project grading	Project grading
14: Jan 13	Project grading