Viterbi decoder on STI CELL processor Michal Blaek - - PowerPoint PPT Presentation

viterbi decoder on sti cell processor
SMART_READER_LITE
LIVE PREVIEW

Viterbi decoder on STI CELL processor Michal Blaek - - PowerPoint PPT Presentation

IBM - CVUT Student Research Projects Viterbi decoder on STI CELL processor Michal Blaek (blazem2@fel.cvut.cz) Viterbi algorithm introduction Finds the most likely state trajectory given the Works in several steps HMM model and


slide-1
SLIDE 1

IBM - CVUT Student Research Projects

Viterbi decoder on STI CELL processor

Michal Blažek (blazem2@fel.cvut.cz)

slide-2
SLIDE 2

Viterbi algorithm introduction

  • Works in several steps
  • Finds the most likely state trajectory given the

HMM model and observation’s sequence.

IBM - CVUT Student Research Projects 2

( ) ( ) ( )

[ ]

( ) ( ) ( )

[ ]

( ) [ ] ( ) [ ]

( )

* 1 1 * 1 * 1 * 1 1 1 1 t 1 1

2,...,1

  • T

1,

  • T

for t ng backtracki . 4 max arg max :

  • n

terminati . 3 max arg max N j 1 T, t 2 for : recursion 2. ) ( N i 1 for : tion initializa 1.

+ + ≤ ≤ ≤ ≤ − ≤ ≤ − ≤ ≤

⋅ = = = = ⋅ = ⋅ ⋅ = ≤ ≤ ≤ ≤ ⋅ = ≤ ≤

t t t T N i T T N i ij t N i t t j ij t N i i i

q q i q i P a i j O b a i j O b i ψ δ δ δ ψ δ δ π δ

slide-3
SLIDE 3

ALF introduction

(Accelerated Library Framework)

  • Programming environment

for data and task parallel applications.

  • Supports the MPMD

(multiple-program-multiple- data) programming model.

  • Optimal tasks scheduling

scheme based on tasks dependencies.

  • Two components: host

runtime X accelerator runtime

  • Input + output +

parameters = work block

IBM - CVUT Student Research Projects 3

slide-4
SLIDE 4

Program realization

  • Algorithm is divide in four separated TASKs creating a simple

pipeline.

  • Between particular neighbouring TASKs is defined a time

dependency for running in right order.

  • TASKs which using ALF are computed in parallel on 1-6 SPEs.
  • Program uses “Host data partitioning” method for partition data

across the accelerators (SPEs).

IBM - CVUT Student Research Projects 4

slide-5
SLIDE 5

Conclusion and results

  • Working correctly.
  • The most difficult part (recursion) is not completed yet

with ALF, for checking results is for now working on PPU(using SIMD) only.

  • No optimalizations and tune.
  • Orientation computation time comparison on the same

data set (HMM-500x500x500, seq-500): – This: 10,5 sec – Old (1SPE, libspe2, SIMD): 18,3 sec

IBM - CVUT Student Research Projects 5