Scientific Computing II (Summer Term 2017)

The lecture treats the efficient implementation of numerical algorithms with special focus on modern hardware architectures. Special implementation techniques for Multicore and GPU machines will be presented with higher insight. Distributed computing basics will also be presented.

Schedule - Certifications - Recommended Literature - Teaching Material - Homework and Exercises - Links

Schedule

Lecture:	on Mondays, 09:15 - 10:45 hrs., G02-106	on Thursdays, 11:00 - 12:30 hrs., G02-210 (even weeks)
Tutorial:	on Thursdays, 11:00 - 12:30 hrs., G02-210 (odd weeks)

Exceptions:

Starting from May 25, we switch the order of Lectures and Excercises in even and odd weeks for the rest of the term.

No lecture on Thursday July 6.

Certifications

Criteria for getting a course assessment:

at least 50% of all possible points from the homework
at least 50% of the points in one homework ( on n-1 of n homework sheets)
demonstrate and explain at least one exercise in the tutorial

Other possibilities to get an assessment will be announced in the lecture.

Recommended Literature

E-Books see Links section

Rauber/Rünger: Parallel Programming; Springer 2010

Hager/Wellein: Introduction to High Performance Computing for Scientists and Engineers; CRC Press 2011

Sanders/Kandrot: CUDA by Example; An Introduction to Genereal-Purpose GPU Programming; Addison Wesley 2011

Teaching Material

Lecture Slides:

Homework and Exercises

Exercise 1 - Apr. 6th, 2017
Exercise 2 - Return at Apr. 24th, 2017 in the lecture or until Apr. 26th, 2017 (12 a.m.) via e-mail.
Exercise 3 - Return at May 8th, 2017 in the lecture or until May 10th, 2017 (12 a.m.) via e-mail.
Exercise 4 - Return at May 31th, 2017 (12 a.m.) via e-mail.
Exercise 5 - June 1th, 2017 - Hands-on Session: Implementation of LU decomposition on top of cuBLAS.