A High-Performance Droplet A High Performance Droplet Routing - - PowerPoint PPT Presentation

A High-Performance Droplet A High Performance Droplet Routing Algorithm for Digital Mi fl idi Bi hi Microfluidic Biochips

Mi ik Ch d D id Z P Minsik Cho and David Z. Pan

• Dept. of Electrical and Computer Engineering

The University of Texas at Austin thyeros@cerc.utexas.edu dpan@ece.utexas.edu http://www.cerc.utexas.edu/utda

1

Sponsored in Part by NSF and IBM Faculty Award

Why Biochip?

Tris-HCL KCL MgCl2 Gelatin Bovine Serum Albumin

Mix Mix

D

Human Experimenter

Mix Mix

Primer Deoxy- nucleotide Triphosphate AmpliTaq DNA λDNA

Mix Mix

p q Polymerase

Mix Mix Mix

Biochip

Result

DNA PCR Biochip

Economical and high-performance

› Low cost (less than $2), portable, disposable ( $ ), p , p › Fast, automated, error-tolerant (no human involvement)

Critical applications

2

› POC (Point-of-care), anti-bioterrorism, …

Digital Microfluidic Biochips

[Courtesy Advanced Liquid Logic]

Ctrl Ckt Ctrl Ckt

P d lt f EWOD

Di iti d d l t t t d

Preprogrammed voltages for EWOD [IEMN]

Digitized droplets transported

› By EWOD (electrowetting-on-dielectric) y ( g )

» Electrical modulation of the solid-liquid interfacial tension

› According to Preprogrammed Schedule

3

g p g

» Traffic control

Microfluidic Biochip Droplet Routing

How to program/schedule/route droplets? › Fluidic constraints to prevent collision (keep-off distance) p ( p ) › Timing constraints to prevent spoilage › NP-Complete

T1 S2 T

2 3 4 5 7 8 9

T1 S2 T

1 2 3

X

2 1 2 6

S T2

1 2

S1 T2 2

3 1

X

2 2 3 4 5 6 7

[Courtesy Advanced Liquid Logic]

S1 S1

T=9 (good solution) T=?? (infeasible) Comparison with VLSI routing Comparison with VLSI routing

› Fluidic constraint = Minimum spacing › Timing constraint = Required arrival time (RAT)

4

› Timing constraint Required arrival time (RAT) › But, time-multiplexed movement of droplets

Modeling and Constraints

t x y

1 x,y,t+1 t+1 x,y,t x,y,t+1 x+1,y,t+1 x,y+1,t+1 1 y+1

(x,y,t)

x-1,y,t+1 x,y-1,t+1

Graph model for simultaneous

x-1 x+1 t-1 y-1 y

p geometric and temporal scheduling Graph model

› 5 edges for each node › Time causality

Fluidic cube

5

› One droplet inside the cube

Current State-of-the Art

Prioritized A* search [Böhringer TCAD’06]

› Route shorter droplets first (widely used in VLSI)

Network flow-based approach [Yuh+ ICCAD’07]

› Maximize the number of nets routed › Min cost-Max flow formulation + prioritized A* search

OSPF protocol approach [Griffith+ TCAD’06]

H t f t d th d h f th b › Have a set of precomputed path, and choose one of them by situation based on OSPF network protocol.

Two-stage Algorithm [Su+ DATE’06] Two stage Algorithm [Su DATE 06]

› Generate M shortest paths › Random selection

Progressive ILP based Approach [Yuh+ DAC’08]

› Similar to VLSI routing

6

› Pin constraints

Proposed Approach

Time-multiplex resource sharing implies

› Intermediate paths will be freed up eventually.

To reduce problem size inspired by Chatin’s

coloring algorithm

New concepts to reflect the nature of biochip

› One droplet movement at a time (the others are frozen)

» Reduced routing search time

› Bypassibility

T t d l t ith i i l i t f ibilit » To route a droplet with minimal impact on feasibility

› Concession

» To resolve a deadlock » To resolve a deadlock

› Compaction

» To satisfy timing constraint and improve fault-tolerance

7

Overall Routing Flow

success

Start

Routing by Bypassibility deadlock? Routing by Concession

failure yes failure success failure no no

Greedy

yes

Unroutable routed?

no

Greedy

• ptimization

yes Reduce the problem size

› To find out the most complex part of the problem › To find out the most complex part of the problem

First find a feasible solution

› Greedily improve the solution to meet timing

8

› Greedily improve the solution to meet timing

One Droplet at a Time

t y

Sj Tj

x y

k

Si Ti Sk Tk

t3

j

Ti Sj Tj Sk

t3

i

Si Ti Tk

t2

Ti Sj Tj Sk Tk

t1

Si Tk

t0

9

Routing by Bypassibility

T1 S2

8 1 2 3 9 4

T

Vleft Vright

T2

1 2 3 4 5 6 7 8 1 2 3 4 5

Hdown

S1

1 2 3 4 5 6

T=9 (optimal solution)

A routed droplet will block the target regions.

I th H/V b f th t d d l t ?

No bypassibility Full bypassibility Full bypassibility

› Is there any H/V bypass for the unrouted droplets?

Four categories

› Ideal: the target is a waste reservoir › Ideal: the target is a waste reservoir. › Full : both horizontal and vertical bypasses are available. › Half : only either horizontal or vertical bypasses is available.

10

› No : no bypass is available.

Routing by Concession

When there is a deadlock…

O d l t d t b k ff › One droplet needs to back-off › One closer to the empty space

11

p y p

Toy Example (Bypassibility) T2 S5 S3 T4 T3 S1 T5 T3 S1 S S S2 S4 T6 T S

12

T1 S6

Toy Example (Bypassibility) T2 S5 S3 T4 T3 T3 S1 T5 T3 T3 S S S1 S S4 T6 S4 T6 S2 T S S

13

T1 S6 S6

Toy Example (Concession) T2 S5 S3 T4 T3 T3 S1 T5

C-Zone

T3 T3 S1 S S

C Zone

S2 S4 T6 T S

14

T1 S6

Toy Example (Bypassibility) T2 S5 S3 T4 T3 S1 T5 T3 S1 S S S2 S4 T6 Operation time = 72 T S

15

T1 S6

Toy Example (Greedy Opt.) T2 S5 S3 T4 T3 T5 S1 T3 S S S1 S2 S4 T6 Operation time = 19 T S

16

T1 S6

Experimental Setup

Implemented in C++ and tested on Intel Dual

Core 2.6GHz Linux with 4GB

Two benchmarks

› Suite1: 4 widely used benchmarks y

» Relatively small and easy » Max # of droplets: 6

S it 2 30 th ti b h k › Suite2: 30 synthetic benchmarks

» Large scale and complex with multiple blockages » Max # of droplets: 48 with 30% of areas blocked » Max # of droplets: 48 with 30% of areas blocked Comparison

› Prioritized A* search [TCAD’06] › Prioritized A search [TCAD 06] › Two-stage algorithm [DATE’06] › Network flow based routing [ICCAD’07]

17

g [ ]

Comparison betw een Suite1 and Suite2

[Yuh+ ICCAD’07]

Much More Complex For Future Design For Future Design

18

Experimental Results

4 P-A* [TCAD'06] 3 P-A [TCAD 06] Two-Stage[DATE'06] NetFlow [ICCAD'07] O 2 Ours

Number of failed designs

1 Suite1

Total 4 designs Completed the most number of designs

19

Completed the most number of designs

Experimental Results

25 P A* [TCAD'06] 20 P-A* [TCAD 06] NetFlow [ICCAD'07] Ours 10 15 Ours

Number of failed designs

5 10

3

Suite2

Total 30 designs Completed the most number of designs

20

Completed the most number of designs

Experimental Results

120 P A* [TCAD'06] 80 100 P-A* [TCAD 06] NetFlow [ICCAD'07] Ours 60 80 Ours

Number of the unrouted

20 40

3

droplets

Suite2

Total 864 droplets (30 designs) Outperforms the previous works by wide margin

21

Outperforms the previous works by wide margin

Conclusion

New digital microfluidic biochip routing algorithm

› Multiple concepts to leverage time multiplexed › Multiple concepts to leverage time multiplexed resource sharing

Outperforms the previous works by wide margin Outperforms the previous works by wide margin Ping-Hung Yuh, Prof. Chia-Lin Yang, and Prof.

Yao Wen Chang (NTU) Yao-Wen Chang (NTU)

› Providing the results of network flow-based on algorithm in ICCAD’07 on benchmark suite2 algorithm in ICCAD 07 on benchmark suite2

Thank you!! Thank you!!

22