 Course description, goals, integrity, knowledge base
 Course number: 2632300, 6 credits
 Spring 2017, lectures: M 10:1512:00, HG D3.2; Th 9:1510:00 CAB G51; occasional substitute lectures: W 13:1515:00 HG D3.2
 Instructor: Markus Püschel
 TAs:
 Alen Stojanov
 Georg Ofenbeck
 Gagandeep Singh
 Only for project supervision: Daniele Spampinato
Grading
 40% research project
 25% midterm
 35% homework
 There is no final Exam
Research Project
# 
Title 
Supervisor/s 
1 
Antinspired edge detection 
GO 
2 
tDistributed stochastic neighbor embedding 
MP 
3 
Marching cubes 
GS 
4 
Locality sensitive hashing 
GS 
5 
PatchMatch algorithm 
GO 
6 
Fractal compression 
AS 
7 
Fast ray tracing for TSDFs 
GS 
8 
Binary convolutional neural network 
MP 
9 
A robust descriptor for line matching 
GO 
10 
Latent Dirichlet Allocation 
GO 
11 
Ray tracing 
GS 
12 
Medial axis transform 
GO 
13 
Quantized neural networks 
AS 
14 
Matrix multiplication over GF(2) 
AS 
15 
Nonlinearly coupled elliptic BFPs 
GO 
16 
GPUCB 
MP 
17 
Online dictionary learning for sparse coding 
MP 
Midterm
26.4, 13:15  15:00, HG E5 (solution, without solution).
Homework
Lectures (including pdfs)
Date 
Content 
Notes 
Other 
20.02 
Course motivation, overview, organization 


23.02 
Cost analysis and performance 


27.02 
Intel Haswell architecture and microarchitecture, memory and computebound 

Intel Haswell, Intel Optimization Manual, Agner Fog’s instruction tables 
01.02 
Instructionlevel parallelism and Compiler limitations 


06.03 
Benchmarking, SIMD (SSE, AVX) overview 


13.03 
SIMD (SSE, AVX) intrinsics 

Intel Intrinsics Guide 
16.03 
SIMD (SSE, AVX) 


22.03 
Locality, caches 


23.03 
Caches, analysis of blocked MMM 
notes 

27.03 
Roofline model 
notes 
paper 
30.03 
Linear algebra libraries, LAPACK, BLAS, ATLAS 


03.04 
Fast MMM (modelbased ATLAS) 
notes 
fast MMM paper 
07.04 
Fast MMM, register renaming 


10.04 
Virtual memory system 
notes 

12.04 
Memory bound computations, sparse MVM 


24.04 
Sparse MVM, linear transforms 


26.06 
Midterm exam 


04.05 
Fast Fourier transform 
notes 

08.05 
Optimizing FFT, FFTW 
notes 
fftw website 
15.05 
Spiral: program generation for transforms 

spiral website 
22.05 
cancelled 


29.05 
Project presentations 


31.05 
Project presentations 


01.06 
Project presentations 

