Presentation is loading. Please wait.

Presentation is loading. Please wait.

Jürgen Lambrecht– DESICS © imec 2002 Multi-objective refinement of Mapping Tables in Telecom Network Applications Jürgen Lambrecht, Chantal Ykman- Couvreur,

Similar presentations


Presentation on theme: "Jürgen Lambrecht– DESICS © imec 2002 Multi-objective refinement of Mapping Tables in Telecom Network Applications Jürgen Lambrecht, Chantal Ykman- Couvreur,"— Presentation transcript:

1

2 Jürgen Lambrecht– DESICS © imec 2002 Multi-objective refinement of Mapping Tables in Telecom Network Applications Jürgen Lambrecht, Chantal Ykman- Couvreur, Armin van der Togt, Francky Catthoor The Protocol Processor Project (PRO 3 ) Contract No IST-1999-11449

3 Jürgen Lambrecht – DESICS © imec 2002 Outline Introduction Context Example: network switch Related Work Our contribution Design space and Method Mapping table implementations: primitives Primitives characteristics Mapping table implementations: multi-layer Exploration Method System Level Cost Factors Results Multi Objective Exploration Impact of network Impact of operation characteristics Future work and Conclusions

4 Jürgen Lambrecht – DESICS © imec 2002 Context Data type refinement Mapping tables Application domain: telecom network, database, multimedia,... Design Technology for Integrated Information and Communication Systems Memory system: bottleneck Methodology: system level specification  optimised embedded implementation

5 Jürgen Lambrecht – DESICS © imec 2002 Example: network switch port 1 Port 2 x25 Y21 Y25 x47 ATM MUX ATM MUX 2 = 0000 0010 5 = 0000 0000 0000 0101 1 = 0000 0000 0000 0001 1 = 0000 0001 4 = 0000 0100 7 = 0000 0000 0000 0111 2 = 0000 0010 5 = 0000 0000 0000 0101 2 = 0000 0010 Key1 ( VPI[8] )Key3 ( port[8] ){ VPI, VCI, port } [32] Key2 ( VCI[16] ) Table (with active connections) 16 284 MB

6 Jürgen Lambrecht – DESICS © imec 2002 Related Work Programming theory STL SETL Embedded systems imec - Wuytack et al. 1996 Design space characterisation for mapping tables Preliminary exploration and optimisation method imec - Ykman et al. ISSS 1999 Extended design space (key merging/splitting) Improved cost function

7 Jürgen Lambrecht – DESICS © imec 2002 Our contribution Area Power Objective trade-off With all relevant application characteristics Automated fast exploration and optimisation simulates the application collects profiling info runs the tool pareto curves result Goal: optimised implementation trade-off multiple objectives taking all relevant system characteristics into account

8 Jürgen Lambrecht – DESICS © imec 2002 Outline Introduction Context Example: network switch Related Work Our contribution Design space and Method Mapping table implementations: primitives Primitives characteristics Mapping table implementations: multi-layer Exploration Method System Level Cost Factors Results Multi Objective Exploration Impact of network Impact of operation characteristics Future work and Conclusions

9 Jürgen Lambrecht – DESICS © imec 2002 Mapping table implementations: primitives key data key data key data Linked List (LL) key data Array (AR) data Pointer Array (PA) data

10 Jürgen Lambrecht – DESICS © imec 2002 Primitives' characteristics

11 Jürgen Lambrecht – DESICS © imec 2002 Mapping table implementations: multi-layer Layer 1 key1 LL(7) Layer 2 PA(4) 2 4 Layer 3 AR(3) data

12 Jürgen Lambrecht – DESICS © imec 2002 Exploration method Key ordering Key hashing Key concatenation Key splitting 2 = 0000 0010 5 = 0000 0000 0000 0101 1 = 0000 0000 0000 0001 1 = 0000 0001 4 = 0000 0100 7 = 0000 0000 0000 0111 2 = 0000 0010 5 = 0000 0000 0000 0101 2 = 0000 0010 Key1 ( VPI[8] )Key3 ( port[8] ){ VPI, VCI, port } [32] Key2 ( VCI[16] ) port 1 Port 2 x25 Y 21 Y25 x47 ATM MUX

13 Jürgen Lambrecht – DESICS © imec 2002 Exploration method Key ordering Key hashing Key concatenation Key splitting Assignment of primitives Cost vector calculation Layer 1 key1 LL(7) Layer 2 PA(4) 2 4 Layer 3 AR(3) data

14 Jürgen Lambrecht – DESICS © imec 2002 System Level Cost Factors Application characteristics Data: size, average number Keys: size, average number, dependencies, frequency distributions Operations: accesses and proportion AR(3) data PA(4) 2 4 Memory cost Area: bits Performance: accesses Power: vendor model

15 Jürgen Lambrecht – DESICS © imec 2002 Outline Introduction Context Example: network switch Related Work Our contribution Design space and Method Mapping table implementations: primitives Primitives characteristics Mapping table implementations: multi-layer Exploration Method System Level Cost Factors Results Multi Objective Exploration Impact of network Impact of operation characteristics Future work and Conclusions

16 Jürgen Lambrecht – DESICS © imec 2002 Multi Objective Exploration Difference Factor 2 in size 5 in power 8 in accesses 2 in performance

17 Jürgen Lambrecht – DESICS © imec 2002 Impact of network

18 Jürgen Lambrecht – DESICS © imec 2002 Impact of operation characteristics

19 Jürgen Lambrecht – DESICS © imec 2002 Future work Better memory library Off-chip memory support On/Off-chip partitioning Include fixed platforms (w. caches…) Further automation New primitive data structures Internet multimedia data sets Trees

20 Jürgen Lambrecht – DESICS © imec 2002 Conclusions Bottleneck in embedded system synthesis: memory performance, area, and power our focus: Mapping Tables Optimisation at system level Automated method needed Multi-objective exploration All relevant application characteristics Fast Power-objective result 90% area savings 80% power savings without decreasing the performance

21 Jürgen Lambrecht – DESICS © imec 2002 www.imec.be Worldwide collaboration with more than 450 companies and institutes. IMEC – Kapeldreef 75 – B-3001 Leuven – Belgium – Tel. +32 16 281211 – Fax +32 16 229400 – www.imec.be

22 Jürgen Lambrecht – DESICS © imec 2002 Dynamic Memory Mngnt. Physical Memory Mngnt. Virtual Memory Segments Selected data structures Physical Memories ADT Abstract Data Type Refinement Virtual memory mngnt Refinement Physical memory mngnt Refinement Memory Management in Matisse

23 Jürgen Lambrecht – DESICS © imec 2002 Exploration and Optimization Power Layers 20,000 2,000 200 100 1,000 10,000 1 2 34


Download ppt "Jürgen Lambrecht– DESICS © imec 2002 Multi-objective refinement of Mapping Tables in Telecom Network Applications Jürgen Lambrecht, Chantal Ykman- Couvreur,"

Similar presentations


Ads by Google