Review

1
Review

• SYSC5603 (ELG6163) Digital Signal Processing
  Microprocessors, Software and Applications
• Miodrag Bolic

2
Lecture 2

• Real-time signal processing
• Stream, block and vector processing
• Parameters
• Throughput
• Range and precision of numbers
• Data-dependent execution,
• Precedence relations within the algorithm,
• Global versus local communication of data
• Random versus regular sequencing of data
  addresses
• What are VLIW and RISC architectures
• What is the difference between CISC, RISC and
  VLIW
• What is Harvard architecture?
• Why the implementation of FIR filter takes a lot
  of time on RISC processors? What is the overhead?

3
Lecture 3

• Understand simple datapaths TMS320C10 and ADSP
  -21x
• Understanding of basic features of DSP
  processors
• Datapath configured for DSP
• Specialized instruction set
• Multiple memory banks and busses
• Specialized addressing modes circular buffering,
  zero-overhead looping
• Specialized peripherals
• Difference between four generations of DSP
  processors
• FIR filtering steps on ADSP -21x
• Basic characteristics of Blackfin processor

4
Lecture 4

• Understand basic assembly functions of TMS320C6x
  PDSPs
• Architecture datapath, instruction packing
• What is linear assembly? What are Intrisic C
  functions?
• What is the code optimization procedure? Describe
  the steps for assembly optimization

5
Lecture 5

• Understanding of Altera FPGAa logic elements and
  Xilinx FPGA CLBs.
• Altera and Xilinx memory blocks
• Altera DSP blocks how one can implement an FIR
  filter using DSP blocks
• Design flow

6
Lecture 6

• Fixed-point representation
• Fixed-point arithmetic
• Simulation methods why does one need range
  estimation? Why the statistics about each
  variable is needed?
• Analytical methods requires knowledge of
  formulae and derivations
• Coefficient quantization
• Overflow quantization in arithmetic operations
• scaling to prevent overflow
• quantization noise statistical modeling
• limit cycle oscillations

7
Lecture 7

• Algorithm Representations and Iteration Bound
• Parallelism and Pipelining
• Retiming
• Unfolding
• Folding

8
Lecture 8

• Signal flow graph in precedence form
• Computational graph for scheduling
• Block scheduling
• Cyclic scheduling
• ASAP, ALAP scheduling

9
Lecture 9

• Convolution
• Fast convolution - Cook-Toom algorithm
• Overlap and save
• FIR filters
• Structures
• Polyphase FIR filters
• Parallel polyphase FIR
• Decimated FIR
• Distributed arithmetic

10
Lecture 10

• FFT Introduction
• Some FFT algorithms
• FFT on PDSP example of the program in C and
  Goertzels algorithm
• FFT floating to fixed-point conversion
• Hardware implementation of FFT pipelined
  inmplementations

11
Lecture 11

• IIR filters
• Pipelining
• Parallelism
• Lattice structure