Parallel Image Processing

1
Parallel Image Processing

• Programming and Architecture

IST PhD Lunch Seminar
Wouter Caarls
Quantitative Imaging Group
2
Why Parallel?

• Processing time
• Smaller timesteps, more scales, faster response
  times
• Memory
• Larger images, more dimensions
• Energy consumption
• More applications, smaller devices

3
Data parallelism

• Many image processing operations have locality
  of reference (segmentation, filtering, distance
  transforms, etc.)
• Data parallelism

4
Task farm parallelism

• An application consists of many different
  operations
• Some of these operations are independent (scale
  spaces, parameter sweeps, noise realizations,
  etc.)
• Task farm parallelism

5
Pipeline parallelism

• An image processing algorithm consists of
  consecutive stages
• If multiple objects are to be processed, they
  may be in different stages at the same time
• Pipeline parallelism

6
Parallel hardware architecturesFine grained

• Irregular
• Superscalar (most modern microprocessors)
• VLIW (DSPs)
• Regular
• Vector (supercomputers, MMX)
• SIMD (graphics processors)
• Custom
• FPGA

7
Parallel hardware architecturesCoarse grained

• Homogeneous
• Multi-core, SMP
• Cluster
• Heterogeneous
• Embedded systems
• Grid

8
Obstacles

• Programming
• Synchronization, bookkeeping
• Different systems, languages, optimization
  strategies
• Choosing an architecture
• Analyze program before it is written
• Additional requirements or unexpected performance
  may require rewrite

9
Architecture-independent parallel programming

• Data parallelism
• Differentiate between synchronization pattern and
  computation
• Library provides pattern, user provides
  computation
• Task farm pipeline parallelism
• Operations do not work on images, but on streams
• Sequences of operation calls do not imply an
  order, but a stream graph.

10
Algorithmic Skeletons
11
Example skeletons

• Pixel
• Neighbourhood
• Recursive neighbourhood

• Stack
• Filter
• Associative reduction

12
Constructing stream graphs
capture
normalize

• By program (dynamic)
• capture(orig)
• normalize(orig, norm)
• dx(orig, x_der, 1.0)
• dy(orig, y_der, 1.0)
• direction(x_der, y_der, dir)
• display(dir)
• Visually (static)

dx
dy
direction
display
13
Mapping stream graphs to processors
14
Dealing with heterogeneous tasks
15
Dealing with interconnect
16
Dealing with dependencies
17
Choosing an architecture automatically

• Architecture-independent program allows automatic
  analyis after it is written, but before an
  architecture is chosen
• Based on certain constraints, architecture can be
  chosen automatically to optimize some cost
  function.
• Tradeoff between cost, power and performance must
  be made by the designer

18
Design Space Exploration
Archi- tecture
Explore
Program
Metrics
Analyze
19
Search strategyConstrained single objective
20
Search strategyMultiobjective tradeoff iteration
21
Search strategyStrength Pareto
22
Conclusions

• Architecture-independent programming allows
• Parallel programming without bookkeeping
• Targeting heterogeneous systems
• Choosing the most appropriate architecture
  automatically
• http//www.qi.tnw.tudelft.nl/wcaarls/smartcam

23
Overview

• Parallelism in image processing
• Parallel hardware architectures
• Architecture-independent parallel programming
• Algorithmic skeletons
• Stream programming
• Choosing an appropriate architecture
• Design Space Exploration

24
Exploiting parallelismFine grained, irregular

• Superscalar
• Dataflow dispatch reorder
• Most modern microprocessors
• Automatic by processor
• Very Long Instruction Word
• Multiple instructions per word
• DSPs, Itanium
• Automatic by compiler

25
Exploiting parallelismFine grained, regular

• Vector instructions
• Supercomputers
• MMX/SSEx
• Special instructions/datatypes
• Single Instruction Multiple Data
• Graphics processors
• Special languages

26
Exploiting parallelismCoarse grained

• Multiprocessing
• Multiple processors/cores sharing a memory
• Shared-memory threading libraries (pthread,
  OpenMP)
• Clusters
• Relatively loosely coupled systems connected by a
  network
• Message-passing libraries (MPI)
• Heterogeneous systems
• Exploit differences in algorithmic requirements
• Multiple paradigms in a single application