Future Directions on Models of Architecture Maxime Pelcat INSA - PowerPoint PPT Presentation

Dataflow workshop - 2017 Future Directions on Models of Architecture Maxime Pelcat INSA Rennes, Institut Pascal

System Design: Y-Chart Application Algorithm Architecture Redesign Redesign Design System Prototype Maxime Pelcat – HDR - 2017 2

System Objectives Reliability Peak Power T ° C Energy Performance Memory $ $ Unit Cost Security Maintenance Cost Maxime Pelcat – HDR - 2017 3

Model-Based Design Model of Architecture (MoA) Model of Computation(MoC) conforms to conform to Algorithm Model Algorithm Architecture Model Redesign Redesign KPI Evaluation KPI Maxime Pelcat – HDR - 2017 4

Models of Architecture Model of Architecture (MoA) conform to Algorithm Model Algorithm Architecture Model Redesign Redesign KPI Evaluation KPI Maxime Pelcat – HDR - 2017 5

MoC is not sufficient Model of Computation(MoC) conforms to Algorithm Model Algorithm Energy Evaluation Energy Maxime Pelcat – HDR - 2017 6

Problem: Predict System Quality • How to predict a DSP system quality ? – Efficiently (simple procedure) – Early (from abstract models) – Accurately (with a good fidelity) – With reproducibility (same models = same prediction) Maxime Pelcat – HDR - 2017 7

Model of Architecture • Definition – Model of a system Non-Functional Property – Application-independent – Abstract – Reproducible Pelcat, M; Mercat, A; Desnos, K; Maggiani, L; Liu, Y; Heulot, J; Nezan, J-F; Hamidouche, W; Ménard, D; Bhattacharyya, S (2017) "Reproducible Evaluation of System Efficiency with a Model of Architecture: From Theory to Practice", IEEE TCAD. Pelcat, M (2018) “ Models of Architecture for DSP Systems", Handbook of Signal Processing Systems, Third Edition, S. S. Bhattacharyya, E. F. Deprettere, R. Leupers , J. Takala, Springer. Maxime Pelcat – HDR - 2017 8

Model of Architecture Model Reproducible Application- Abstract independent    AADL    MCA SHIM   /   UML MARTE    AAA    CHARMED    S-LAM    MAPS    LSLA Maxime Pelcat – HDR - 2017 9

Model of Architecture Model G conforms to MoC Model H conforms to MoA Activity MoA depends on MoC One and always the same quality evaluation Reliability Power NFP = MoA( ) activity( ) MoC( ) application Energy Performance Memory Security Cost T ° C Maxime Pelcat – HDR - 2017 10

Model of Architecture MoC MoA Act KPI Maxime Pelcat – HDR - 2017 11

LSLA: First MoA • LSLA = Linear System-Level Architecture Model • Motivated by the additive nature of energy consumption Maxime Pelcat – HDR - 2017 12

LSLA Model of Architecture Task3 Task2 1 1 1 signal signal Task5 Task1 1 1 1 1 Task4 token quantum 16+12+22=50 2x+0 3x+0 PE1 PE2 CN Compositional Maxime Pelcat – HDR - 2017 10x+1 13

LSLA Model of Architecture Task3 Task2 1 1 1 signal signal Task5 SDF: Model of Computation Task1 1 1 1 1 Task4 Activity 16+12+22=50 2x+0 3x+0 PE1 PE2 CN LSLA: Model of Architecture Maxime Pelcat – HDR - 2017 10x+1 14

LSLA MoA for Energy Prediction • 86% of fidelity on octo-core ARM  Maxime Pelcat – HDR - 2017 15

LSLA MoA for Energy Prediction • The model is learnt from energy measurements PE PE CN PE PE CN PE PE CN PE PE Maxime Pelcat – HDR - 2017 16

LSLA MoA for Energy Prediction • The model is learnt from energy measurements 1.5W 1.5W PE PE α CN 1.5W 1.5W PE PE CN 0.3W 0.3W PE PE β CN 0.3W 0.3w γ PE PE Maxime Pelcat – HDR - 2017 17

MoAs: Limits of LSLA  • Energy  Linear model OK • Latency ! • Latency does not have an additive nature 1 Latency = sum Task1 Task2 1 1 Task1 1 1 Latency = max 1 Task2 1 Maxime Pelcat – HDR - 2017 18

Activity & MoA for Latency SDF Task3 Task2 1 1 1 signal signal Task5 Task1 1 1 1 1 Task4 a) b) Maxime Pelcat – HDR - 2017 19

Activity & MoA for Latency Σ  12+12+11=35 a) b) Σ  8+6+11=25 max(35,25)=35 2x+0 3x+0 PE1 PE2 CN MaxPlus Maxime Pelcat – HDR - 2017 10x+1 20

Accuracy?? No! Fidelity!! Application Algorithm Architecture Redesign Redesign Design System Prototype Maxime Pelcat – HDR - 2017 21

Directions for Research on MoA MoC MoA • Try existing models on new KPIs Act • Create new models for new KPIs – When existing ones do not match KPI • Co-explore MoAs - multi-objective optim. • Learn more complex, non-linear models Maxime Pelcat – HDR - 2017 22