MI L

FPGA Co-Processor Enhanced Ant Colony Systems Data Mining

Jason Isaacs and Simon Y. Foo
Machine Intelligence Laboratory FAMU-FSU College of Engineering
Department of Electrical and Computer Engineering

MAPLD2005/P249


Isaacs

Presentation Outline Introduction Significance of Research Concise Background on ACS Summary of Data Mining focused on Clustering Discussion of ACS-based Data Mining FPGA Co-processor Enhancement Conclusions Future Work
2

MI L

MAPLD2005/P249

MI L
Project Goal: to design and implement an Ant Colony
Systems toolbox for non-combinatorial problem solving. This toolbox will comprise both hardware and software based solutions.

Isaacs

MAPLD2005/P249

MI Ant Colony Systems Project Overview L

This work aims at advancing fundamental research in Ant Colony Systems. The major objectives of this project are:
Develop a set of behavior models Design ACS algorithms for solutions to non-combinatorial problems Analyze algorithms for hardware implementations Implement FPGA Modules CURRENT Incorporate all modules into a cohesive toolbox

Isaacs

MAPLD2005/P249

MI Introduction to Ant Colony Systems L

Ants are model organisms for bio-simulations due to both their relative individual simplicity and their complex group behaviors. Colonies have evolved means for collectively performing tasks that are far beyond the capacities of individual ants. They do so without direct communication or centralized control Stigmergy. Previous Research: our use of simulated ants to generate random numbers proved a novel application for ACS.
Prior to 1992, ACS was used exclusively to study real ant behavior. However, in the last decade, beginning with Marco Dorigos 1992 PhD Dissertation Optimization, Learning and Natural Algorithms, modeling the way real ants solve problems using pheromones, ant colony simulations have provided solutions to a variety of NP-hard combinatorial optimization problems

Isaacs

MAPLD2005/P249

MI L ACS Application Area: Data Mining

Ant Colony real-world behaviors applicable to Data Mining:
Ant Foraging Cemetery Organization and Brood Sorting Division of Labor and Task Allocation Self-organization and Templates Co-operative Transport Nest Building

Isaacs

MAPLD2005/P249

MI Cemetery Organization and Brood Sorting L

Isaacs

MAPLD2005/P249

Ant Colony Nest Examples

MI L

Isaacs

MAPLD2005/P249

MI Flowchart for the ACS Data Mining System L

Data Feature/Object Update Cognitive Map NO YES ACS Data Mining

Store New Object

Recognize NEST (Data Warehouse)

Classification Clustering Connection Topology

Isaacs

MAPLD2005/P249

Knowledge Discovery and Data Mining

What is Data Mining?

MI L

Discovery of useful summaries of data Also, Data Mining refers to a collection of techniques for extracting interesting relationships and knowledge hidden in data. It is best described as the nontrivial process of identifying valid, novel, potentially useful, and ultimately understandable patterns in data. (Fayyad, et al 1996)

Isaacs

10

MAPLD2005/P249

MI Knowledge Discovery in Databases L

Cleaning Integration Selection Transformation Data Mining Evaluation Visualization

Data Warehouse

Prepared data Patterns Knowledge Knowledge Base

Data

Isaacs

11

MAPLD2005/P249

MI Typical Tasks in Data Mining L

Classification Prediction Clustering Association Analysis Summarization

Isaacs

12

MAPLD2005/P249

Clustering What is Clustering?

MI L

Given points in some space, often a highdimensional space, group the points into a small number of clusters, each cluster consisting of points that are near in some sense.

Isaacs

13

MAPLD2005/P249

MI The k-Means Algorithm L the k-means picks k cluster centroids and assigns points to the clusters by picking
closest centroid to the point in question. As points are assigned to clusters, the centroid of the cluster may migrate.

For a very simple example of five points in two dimensions. Suppose we assign the points 1, 2, 3, 4, and 5 in that order, with k = 2. Then the points 1 and 2 are assigned to the two clusters, and become their centroids for the moment. When we consider point 3, suppose it is closer to 1, so 3 joins the cluster of 1, whose centroid moves to the point indicated as a. Suppose that when we assign 4, we find that 4 is closer to 2 than to a, so 4 joins 2 in its cluster, whose center thus moves to b. Finally, 5 is closer to a than to b, so it joins the cluster {1,3}, whose centroid moves to c.

Isaacs

14

MAPLD2005/P249

The k-Means Algorithm

MI L

Having located the centroids of the k clusters, we can reassign all points, since some points that were assigned early may actually wind up closer to another centroid, as the centroids move about. If we are not sure of k, we can try different values of k until we find the smallest k such that increasing k does not much decrease the average distance of points to their centroids.

Isaacs

15

MAPLD2005/P249

ACS Notation and Heuristics

E = {Oi,, On} Set of n data or objects collected.

MI L

Oi = {vi,, vk} Each object is a vector of k numerical attributes. Vector similarity is measured by Euclidean distance (can use other: Minkowski, Hamming, or Mahalanobis). Dmax = max D{Oi, Oj}, where Oi,Oj E

Isaacs

16

MAPLD2005/P249

MI ACS Notation and Heuristics L

2-D search area, in general, must be at least m2 n, but experiments have shown that m2 4n provides good results. A heap/pile H is considered to be a collection of two or more objects. This collection is located on a given single cell rather than just spatially connected. This limitation prevents overlaps.
O1 O3 O2 O4 O1 O5 O3 O2 O4 O5

Spatial pattern cluster

Isaacs 17

Single-cell ranked cluster

MAPLD2005/P249

ACS Distance Measures

Dmax is the maximum distance between two objects of H:
Dmax ( H ) = max D(Oi , O j )
O i ,O j H

MI L

Ocenter is the center of mass of all objects in H: (not necessarily a real object) 1
Ocenter ( H ) = nH
O i H

Oi

Odissim is the most dissimilar object in H, i.e. which maximizes

D(., Ocenter ( H ))

Dmean is the mean distance between the objects of H and the center of 1 mass Ocenter : Dmean ( H ) = D(Oi , Ocenter ( H ))
nH
O i H

Isaacs

18

MAPLD2005/P249

MI ACS Unsupervised Learning and Clustering L Algorithm

Initialize randomly the ant positions Repeat For each anti Do
Move anti If anti does not carry any object Then look at 8-cell neighborhood and pick up object according to pick-up algorithm Else (anti is already carrying an object O) look at 8-cell neighborhood and drop O according to drop-off algorithm

Until stopping criterion

Isaacs

19

MAPLD2005/P249

ACS Data Mining Algorithm Top Level

1. 2. 3. 4. 5. Load Database Data Compression Object Clustering Clustering of Similar Groups Reevaluate Objects in Groups

MI L

Isaacs

20

MAPLD2005/P249

ACS Data Mining Algorithm Top Level

Load Database Select Compression Method
Wavelets Principle Component Analysis None

MI L

Repeat for Max_Iterations1 Object Clustering

Begin Ants Redistribute Objects K-means

Repeat for Max_Iterations2 Clustering of Similar Groups

Ants Redistribute Piles (Clusters) of Objects K-means

Repeat for Max_Iterations3 Reevaluate Objects in Groups

Ants Redistribute Objects in Clusters with a Probability based on Least Similar Objects Distance from the Mean of the Cluster K-means

Isaacs

21

MAPLD2005/P249

MI ACS Object Pick-up AlgorithmL

1. 2. Label 8-cell neighborhood as unexplored Repeat
1. 2. Consider the next unexplored cell c around anti with the following order: cell 1is NW, cell 2 is N, cell 3 is NE, N is the direction the ant is facing. If c is not empty Then do one of the following:
1. 2. 3. If c contains a single object O, Then load O with probability Pload, Else If c contains a heap of two objects, Then remove one of the two with a probability Pdestroy, Else If c contains a heap H of more than 2 objects, Then remove the most dissimilar object Odissim(H) from H provided that

3.

Label c as explored

D (Odissim ( H ), Ocenter ( H )) > Tremove Dmean ( H )

3.

Until all 8 cells have been explored or one object has been loaded

Isaacs

22

MAPLD2005/P249

MI ACS Object Drop-off Algorithm L

1. 2. Label 8-cell neighborhood as unexplored Repeat
1. Consider the next unexplored cell c around anti with the following order: cell 1is NW, cell 2 is N, cell 3 is NE, N is the direction the ant is facing.
1. 2. If c is empty Then drop O in cell with a probability Pdrop, Else If c contains a single object O, Then drop O to create a heap H provided that: D (O, O ' ) > Tcreate Dmax Else If c contains a heap H, Then drop O on H provided that:

3.

D(O, Ocenter ( H )) < D(Odissim ( H ), Ocenter ( H ))

2.

Label c as explored

3.

Until all 8 cells have been explored or carried object has been dropped

Isaacs

23

MAPLD2005/P249

Parameter Table
Parameter Speed Pdirection Maxcarry Pload Pdestroy Pdrop Tremove Tcreate Role Distance ant can travel in one time step Probability to move in same direction Maximum object carry time Probability to pick-up an object Probability to destroy an heap of 2 objects Probability to drop an object Minimum dissimilarity for removing an object from a heap Maximum dissimilarity permitted for creating a heap of two objects
24

MI L Value (or Range)

[1,10] [0.5,1] [20,200] [0.4,0.8] [0,0.6] [0.4,0.8] [0.1,0.2] [0.05,0.2]

Isaacs

MAPLD2005/P249

K-means Algorithm

MI L

1. Take as input the partition P of the data set found by the ants in the form of k heaps: Hi,,Hk 2. Repeat
1. Compute Ocenter(Hi),, Ocenter(Hk) 2. Remove all objects from heaps, 3. For each object Oi E:
1. Let Hi, j [1, k] be the heap whose center is the closest to Oi, 2. Assign Oi to Hj,

4. Compute the resulting new partition P = H1,,Hk by removing all empty clusters,

3. Until stopping criterion

Isaacs 25 MAPLD2005/P249

Benchmark Databases

MI L

The following public domain data sets were obtained from the UCI (University of California at Irvine) Machine Learning Repository. These have been used extensively for classification tasks using different paradigms. The main characteristics of each of these domains are described in the three slides.

Isaacs

26

MAPLD2005/P249

Tested Databases
Golf
Very simple database, 4 attributes, 2 classes

MI L

Balloons
The influence of prior knowledge on concept acquisition, 4 data sets, 4 attributes, 2 classes

Wine
Well behaved class structure, 178 instances, 13 attributes, 3 classes

Hepatitis
Poorly distributed database, 155 instances, 19 attributes, 2 classes

Iris (plant)
Very popular database, 150 instances, 4 attributes, 3 classes.

Wisconsin Breast Cancer

High dimensional database, 198 instances, 32 attributes, 2 classes

Isaacs

27

MAPLD2005/P249

Golf Data Results

Outlook Temperature Humidity Windy Decision sunny continuous continuous true/false play/don't play overcast rain

MI L

sunny sunny overcast rain rain rain overcast sunny sunny rain sunny overcast overcast rain

85 80 83 70 68 65 64 72 69 75 75 72 81 71

85 90 78 96 80 70 65 95 70 80 70 90 75 80

FALSE TRUE FALSE FALSE FALSE TRUE TRUE FALSE FALSE FALSE TRUE TRUE FALSE TRUE

Dont' Play Dont' Play Play Play Play Dont' Play Play Dont' Play Play Play Play Play Play Dont' Play

1 1 2 3 3 3 2 1 1 3 1 2 2 3

85 80 83 70 68 65 64 72 69 75 75 72 81 71

85 90 78 96 80 70 65 95 70 80 70 90 75 80

0 1 0 0 0 1 1 0 0 0 1 1 0 1

0 0 1 1 1 0 1 0 1 1 1 1 1 0

0.3333 0.3333 0.6667 1 1 1 0.6667 0.3333 0.3333 1 0.3333 0.6667 0.6667 1

1 0.9412 0.9765 0.8235 0.8 0.7647 0.7529 0.8471 0.8118 0.8824 0.8824 0.8471 0.9529 0.8353

0.8854 0.9375 0.8125 1 0.8333 0.7292 0.6771 0.9896 0.7292 0.8333 0.7292 0.9375 0.7813 0.8333

0 1 0 0 0 1 1 0 0 0 1 1 0 1

0 0 1 1 1 0 1 0 1 1 1 1 1 0

Given Data

Numerical Equivalent

Normalized

Isaacs

28

MAPLD2005/P249

Golf Data Results

MI L

Number in Cluster

ERROR
3 4 2 1 3 4 0 0 1 1 0 0 0 1 0 0 1 0 0 1 0 1 0 0 0 1 0 0 0 1 0 1 0 0 1 0 1 0 0 0 1 0 0 1 0 1 0 0

Dont Play Play Dont Play

4 8 2

Position of Cluster

Objects (1-14)

Isaacs

29

MAPLD2005/P249

Golf Data Results

MI L

Number in Cluster

No Errors
Play Dont Play Dont Play

9 3 2

3 4 5

3 3 5

0 0 1

0 1 0

1 0 0

1 0 0

1 0 0

0 1 0

1 0 0

0 0 1

1 0 0

1 0 0

1 0 0

1 0 0

1 0 0

0 1 0

Position of Cluster

Objects (1-14)

Isaacs

30

MAPLD2005/P249

MI Wine Database L Data is the results of a chemical analysis of wines grown in the same region
in Italy but derived from three different cultivars.
The attributes are 1) Alcohol 2) Malic acid 3) Ash 4) Alcalinity of ash 5) Magnesium 6) Total phenols 7) Flavanoids 8) Nonflavanoid phenols 9) Proanthocyanins 10)Color intensity 11)Hue 12)OD280/OD315 of diluted wines 13)Proline

Number of Instances class 1 59 class 2 71 class 3 48

Error: 0.050562 5 class 1 mislabeled as class 2 3 class 2 mislabeled as class 3 1 class 3 mislabeled as class 2

Isaacs

31

MAPLD2005/P249

MI Iris (Plant) Database L This is perhaps the best known database to be found in the pattern recognition literature.
Attribute Information: 1. sepal length in cm 2. sepal width in cm 3. petal length in cm 4. petal width in cm
Errors:4 mislabeled as #2 Errors:3 mislabeled as #3 0.046667

Number of Instances: 150 (50 in each of three classes) -- Iris Setosa -- Iris Versicolour -- Iris Virginica

Errors:2 mislabeled as #3 Errors:4 mislabeled as #2 0.04

Errors: 0.047 4 mislabeled as type 2 3 mislabeled as type 3

Errors: 0.04 2 mislabeled as type 3 4 mislabeled as type 2

Isaacs

32

MAPLD2005/P249

 Number of Total Iterations Compression Method (PCA, Wavelet, None) Cluster Method Objects Only Objects and Groups of Objects Objects, Groups, then Objects again Number of Ants K-Means Iterations Distance Measure (Euclidean, Minkowski, Hamming, or Mahalanobis) Others (RNG, Ants Movement Distance, Ant Carrying Capacity)
Isaacs 33 MAPLD2005/P249

MI ACS DM: Optimization of Parameters L

MI ACS DM: Object Grouping Only L

Database WPBC WPBC WPBC WPBC WPBC WPBC Wine Wine Wine Wine Wine Wine Hepatitis Hepatitis Hepatitis Hepatitis Hepatitis Hepatitis Iris Iris Iris Iris Iris Iris Compression None Wavelet DB3 PCA None Wavelet DB3 PCA None Wavelet DB3 PCA None Wavelet DB3 PCA None Wavelet DB3 PCA None Wavelet DB3 PCA None Wavelet DB3 PCA None Wavelet DB3 PCA Vector Length 33 12 9 33 12 9 13 7 7 13 7 7 19 8 13 19 8 13 4 4 1 4 4 1 Hybrid Number of Number of Iterations Ant Count Max_ACS Max_K-Means Data Points Resultant Groups 50 80 30 15 198 34 50 80 30 15 198 29 50 80 30 15 198 47 100 80 30 15 198 32 100 80 30 15 198 18 100 80 30 15 198 34 50 80 30 15 178 36 50 80 30 15 178 25 50 80 30 15 178 33 100 80 30 15 178 20 100 80 30 15 178 22 100 80 30 15 178 23 50 80 30 15 80 22 50 80 30 15 80 18 50 80 30 15 80 24 100 80 30 15 80 18 100 80 30 15 80 16 100 80 30 15 80 14 50 80 30 15 150 5 50 80 30 15 150 14 50 80 30 15 150 6 100 80 30 15 150 5 100 80 30 15 150 14 100 80 30 15 150 6 Error Num ADM 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1

1.6854 7.3034 1.6854 2.809 7.3034 1.6854 8.75 13.75 11.25 8.75 16.25 13.75 4 3.3333 6 4 3.3333 6

Isaacs

34

MAPLD2005/P249

MI ACS DM: Object and Cluster Grouping Only L

Database WPBC WPBC WPBC WPBC WPBC WPBC Wine Wine Wine Wine Wine Wine Hepatitis Hepatitis Hepatitis Hepatitis Hepatitis Hepatitis Iris Iris Iris Iris Iris Iris Compression None Wavelet DB3 PCA None Wavelet DB3 PCA None Wavelet DB3 PCA None Wavelet DB3 PCA None Wavelet DB3 PCA None Wavelet DB3 PCA None Wavelet DB3 PCA None Wavelet DB3 PCA Vector Length 33 12 9 33 12 9 13 7 7 13 7 7 19 8 13 19 8 13 4 4 1 4 4 1 Hybrid Number of Number of Iterations Ant Count Max_ACS Max_K-Means Data Points Resultant Groups 50 80 30 15 198 21 50 80 30 15 198 18 50 80 30 15 198 43 100 80 30 15 198 13 100 80 30 15 198 16 100 80 30 15 198 29 50 80 30 15 178 33 50 80 30 15 178 21 50 80 30 15 178 30 100 80 30 15 178 19 100 80 30 15 178 14 100 80 30 15 178 21 50 80 30 15 80 22 50 80 30 15 80 17 50 80 30 15 80 24 100 80 30 15 80 18 100 80 30 15 80 11 100 80 30 15 80 14 50 80 30 15 150 4 50 80 30 15 150 4 50 80 30 15 150 3 100 80 30 15 150 3 100 80 30 15 150 6 100 80 30 15 150 5 Error Num ADM 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2

1.6854 6.7416 1.6854 2.809 8.9888 1.6854 8.75 13.75 11.25 8.75 16.25 13.75 13.333 13.333 3.333 4 8 6

Isaacs

35

MAPLD2005/P249

MI ACS DM: Object, Cluster, and Object L

Database WPBC WPBC WPBC WPBC WPBC WPBC Wine Wine Wine Wine Wine Wine Hepatitis Hepatitis Hepatitis Hepatitis Hepatitis Hepatitis Iris Iris Iris Iris Iris Iris Compression None Wavelet DB3 PCA None Wavelet DB3 PCA None Wavelet DB3 PCA None Wavelet DB3 PCA None Wavelet DB3 PCA None Wavelet DB3 PCA None Wavelet DB3 PCA None Wavelet DB3 PCA Vector Length 33 12 9 33 12 9 13 7 7 13 7 7 19 8 13 19 8 13 4 4 1 4 4 1 Hybrid Number of Number of Iterations Ant Count Max_ACS Max_K-Means Data Points Resultant Groups 50 80 30 15 198 19 50 80 30 15 198 18 50 80 30 15 198 38 100 80 30 15 198 12 100 80 30 15 198 14 100 80 30 15 198 27 50 80 30 15 178 31 50 80 30 15 178 21 50 80 30 15 178 29 100 80 30 15 178 7 100 80 30 15 178 11 100 80 30 15 178 21 50 80 30 15 80 22 50 80 30 15 80 16 50 80 30 15 80 22 100 80 30 15 80 16 100 80 30 15 80 8 100 80 30 15 80 11 50 80 30 15 150 3 50 80 30 15 150 4 50 80 30 15 150 3 100 80 30 15 150 3 100 80 30 15 150 6 100 80 30 15 150 5 Error Num ADM 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3

1.6854 6.7416 1.6854 2.2472 6.7416 2.2472 8.75 12.5 11.25 10 16.25 16.25 4 13.333 3.3333 4 8 6

Isaacs

36

MAPLD2005/P249

Why Move to Hardware?

MI L

For such large datasets the ACS classifier perform remarkably well. However,
Speed of classification is very limited in software. The computational bottlenecks lay in the number of multiply and adds that must be performed for each object. In addition, the requirement of a square root for each distance measurement adds complexity.

Isaacs

37

MAPLD2005/P249

Target Hardware:
Avnets Virtex II Pro Board

MI L

Uses Virtex II Pro XC2VP20 Many Options for I/O. 32 Bit PCI Bus has Data Throughput of Over 100 MB per Second.

Isaacs

38

MAPLD2005/P249

MI ACS-DM System Top-Level HW L

ACS Cellular Automata Random Number Generator Ant Colony Actions and Data Module K-Means Data Actions and Information

Database Actions and Information Module

Data Comparison Module

Isaacs

39

MAPLD2005/P249

ACSDM Hardware Design

MI L

Isaacs

40

MAPLD2005/P249

MI K-Means Distance Calculator with L

CORDIC Square Root

Isaacs

41

MAPLD2005/P249

Device Utilization Summary

Selected Device : 2vp20ff896-6 Number of Slices: 6600 out of 9280 71% Number of Slice Flip Flops: 8312 out of 18560 44% Number of 4 input LUTs: 7661 out of 18560 41% Number of bonded IOBs: 266 out of 556 48% Number of BRAMs: 3 out of 88 3% Number of MULT18X18s: 8 out of 88 9% Number of GCLKs: 1 out of 16 6% ========================================================================= TIMING REPORT Clock Information: -----------------------------------+------------------------+-------+ Clock Signal | Clock buffer(FF name) | Load | -----------------------------------+------------------------+-------+ clk | BUFGP | 1419 | -----------------------------------+------------------------+-------+ Timing Summary: Minimum period: 16.499ns (Maximum Frequency: 60.611MHz) Minimum input arrival time before clock: 4.491ns Maximum output required time after clock: 6.087ns Maximum combinational path delay: 5.102ns

MI L

CORDIC Sqrt data path is greatest bottleneck causing high period

Isaacs

42

MAPLD2005/P249

MI Hardware Euclidean Distance Result L

V1
0.49655 0.89977 0.82163 0.64491 0.81797 0.66023 0.34197 0.28973 0.34119 0.53408 0.72711 0.30929 0.8385 0.56807 0.37041 0.70274 0.54657 0.44488 0.69457 0.62131

V2

0.83812 0.01964 0.68128 0.37948 0.8318 0.50281 0.70947 0.42889 0.30462 0.18965 0.19343 0.68222 0.30276 0.54167 0.15087 0.6979 0.37837 0.86001 0.85366 0.59356

Dist =

V1
n =1

V 2n

Result from Matlab = 1.5058 Result from Hardware = 1.5172 Vectors are Fix 8_7 on input
Then after add: Fix 9_7 Then after multi: Fix 18_14 Then after accum: Fix 20_14 Then after CORDIC Sqrt: Fix 42_36

Error is present in round-off and Cordic Sqrt

Isaacs

43

MAPLD2005/P249

MI Ant Colony Actions: MovementL

CARNG is a simple 32-bit rule 30 that is user initialized for reproducibility
RNG Ant(1) RNG Ant(2)

Ant Move-Direction Filter

Current Location Data

Ant Change Location

Ant Colony Data

New Location Data

Current Location Last Location Have Data Status

RNG Ant(N)

Isaacs

44

MAPLD2005/P249

Pheromone Trail Result from Hardware Co-simlulation

MI L

A single ant is simulated for clarity and the Darker Red is most recent position

Isaacs

45

MAPLD2005/P249

Ant Colony Actions: Object Load/Drop

Enable Drop/Load Y/N
Current Location

MI L

Were Probabilities and Thresholds Met?

Carried Status

Object Information
Current Have Data Status

Current Location Carried Status

Ant Change Have Data Status

Ant Colony Data

New Have Data Status

Current Location Last Location Have Data Status

Isaacs

46

MAPLD2005/P249

MI ACS DM Hardware: Storage Requirements L

Preprocessed Data (Number of Objects * Vector Length, 8- to 32-bit fixed-point)
Object Vectors Object Locations Object Status

Parameter Values (16 32-bit fixed-point)

Probabilities Thresholds Limits

Max Distance (1 32-bit fixed-point) Groups (Number of Objects * Number of Groups, 1-bit and 3*Number of Groups 8-bit)
Members Means (Object Vector Length * 32-bit fixed-point) Locations

Ant Locations and Have-Object Status (Number of Ants * 8-bit, plus 1-bit status)

Isaacs

47

MAPLD2005/P249

MI L

Isaacs

48

MAPLD2005/P249

PCI Bridge

MI L

Isaacs

49

MAPLD2005/P249

Block Diagram

MI L

Virtex-II Pro is focal point. Spartan acts as bridge to PCI On Board Memory 32 MB SDRAM 2 MB SRAM 16 MB FLASH 128 MB DDR SDRAM 64 MB Compact Flash Ethernet RS232 4 AvBus Connectors 2 PMC Connectors
Isaacs 50 MAPLD2005/P249

Conclusions/Future Work
Continue to design the ACS Data Mining System
Implement an improved Memory Manager Correct Errors associated with Round-off and the CORDIC Sqrt. Implement the Group Clustering Algorithm

MI L

Optimize the PC/FPGA interfacing to create our own low-cost integrated system.
Our problems currently reside on the PCI interface design shipped with the Avnet Development Board. We are working hard to resolve this issue, but in the end we may have to consider another board. Also shown in presentation P248.

We also need to improve the speed. 60Mhz is too slow. Optimize data through put and calculating efficiency of the distance metric algorithm, i.e., consider a multi-stage pipeline or employ the use of more look-up tables. The ultimate goal is to demonstrate the ability of ACS algorithms to perform as well as other well-know techniques allowing for computational speed-up utilizing FPGAs as co-processors.

Isaacs

51

MAPLD2005/P249