Download presentation
Presentation is loading. Please wait.
Published byMoris Lester Modified over 9 years ago
1
Alternative Parallel Processing Approaches Jonathan Sagabaen
2
Overview Neural Networks Systolic Arrays Dataflow Computing
3
Data flows continuously and is available to multiple instructions at the same time Each instruction is considered to be a separate process
4
An instruction is executed when the data necessary for execution become available Example >> Instructions do not reference memory; instead, they reference other instructions Dataflow Computing
5
Updating spreadsheet In a spreadsheet you can specify a cell formula which depends on other cells Then when any of those cells is updated the first cell's value is automatically recalculated
6
Neural Networks Also known as Connectionist systems Adaptive systems Parallel distributed processing systems Neural network computers are composed of a large number of simple processing elements that individually handle one piece of a much larger problem
7
Neural Networks A neural net consists of processing elements (PEs), which multiply inputs by various sets of weights, yielding a single output value As a neural network is learning, it can make a mistake As a neural network is learning, it can make a mistake When it does, weights and thresholds must be changed to compensate for the error Example >>
8
Chicken example You show a child a chicken and tell him it's a bird Then you show the child an owl However, if we show a child a chicken, a sparrow, a duck, a hawk, a pelican, and a crow, then it's more likely he'll make the conclusion that it is a bird. It's likely that the child won't recognize the owl as a bird
9
Commonly used for Quality control Financial and economic forecasting Speech and pattern recognition Health care cost reduction Bankruptcy prediction Machine Diagnostics Securities trading Target Marketing
10
Commonly used for Artificial Intelligence Weather forecasting Oil and gas exploration
11
Systolic Arrays They are a network of processing elements that rhythmically compute data by circulating it through the system A systolic array is a pipe network arrangement of processing units called cells where each cell compute data and store it independently of one another
12
Systolic Arrays ProsCons Highly specialized thus inflexible to problems to solve Short connections Simple design Highly scalable Robust Compact Efficient Cheap
13
Commonly used for repetitive tasks such as: Fourier transformations Image processing Data compression Shortest path problems Sorting Signal processing Matrix computations
Similar presentations
© 2025 SlidePlayer.com. Inc.
All rights reserved.