An introduction to Esterel and its compilation

1. An introduction to Esterel and its compilation Dumitru Potop, fresh PhD

2. Outline • (Long) Introduction • Embedded systems, Reactive systems, Synchronous paradigm • Esterel • Syntax and intuitive semantics • Esterel translation schemes • Automata-based • Circuit-based • “Simulation”-based

3. Embedded systems • Airplanes, cars, phones, washing machines, video cameras contain computers! • Complex, custom hardware/software architectures • Specific problems in their development: • Specification, due to interdisciplinarity • Specific languages, methods • Compilation • Implementations must fit tight specifications • Joint development of hardware and software • Insuring correctness, fault tolerance, and robustness (expensive bugs)

4. Reactive systems (Harel, 1987) • Emphasis on communication and control • React to input events with appropriate output events, according to the state • Examples: digital circuits, man-machine interfaces, communication protocols state … Inputs Outputs …

5. Synchronous paradigm • Cycle-based execution model, global clock • Perfect synchrony • Causality Input event Output event computation memory

6. Synchronous paradigm • Synchrony = abstraction of the real world: • Natural in circuits • Advantages: clear semantics, codesign, verification methods • Problem: combining circuit semantics with a higher-level description style time Execution instants

7. Synchronous languages • Esterel (Berry et al.), ECL • Structural imperative (Ada-like) style • Fit for control-dominated parts • Lustre (Halbwachs et al.), Signal (LeGuernic et al) • Dataflow networks • Circuit-like, low-level • Argos (Maranichi), SyncCharts (André) • Visual Statecharts-like formalisms

8. The Esterel language • Structural imperative style • Intuitive, modular • Basic constructs • Classical control flow: p;q,p||q, loop p end • Preemption:abort p when S, suspend p when S • Exception handling: trap T in p end, exit T • Division into instants: pause, await S • Signals: signal S in p end, emit S, present S then p else q end

9. The Esterel language • Example 1: loop [ await A || await B ]; emit O; halt every R

10. The Esterel language • Example 2: signal S in await I ; emit S || await S ; emit O end signal

11. The Esterel language • Constructive causality • Correct causality cycles • Instantaneous reaction to signal absence (analysis of not yet executed code) causality cycle signal S,T in emit S; present T then present S else emit T end end; end break the cycle

12. Old translation schemes 0 • Automata translation • Fast, large code 1 2 3 loop [ await A || await B ]; emit O; halt every R 4 switch(state){ case 0: state=1; break; case 1: if(!R)if(A)if(B) {O=1;state=4;} else state=2; else if(B)state=3;break; case 2: if(R)state=1; else if(B){O=1;state=4;} break; case 3: if(R)state=1; else if(A){O=1;state=4;} break; case 4: if(R)state=1;break; }

13. Old translation schemes A • Circuit translation • Slow, small code R O loop [ await A || await B ]; emit O; halt every R boot B RESET = R & !BOOT A_TRIGGER = A_OUT & !RESET A_THEN = A_TRIGGER & A A_ELSE = A_TRIGGER & !A A_TERM = A_THEN | ! A_TRIGGER A_IN = BOOT | RESET | A_ELSE B_TRIGGER = B_OUT & !RESET B_THEN = B_TRIGGER & B B_ELSE = B_TRIGGER & !B B_TERM = B_THEN | ! B_TRIGGER B_IN = BOOT | RESET | B_ELSE ACTIVE = A_TRIGGER | B_TRIGGER O = A_TERM & B_TERM & ACTIVE

14. Old translation schemes • Direct semantic simulation (Edwards, Bertin) • Fast, small code through static scheduling • Restrict the class of accepted programs Resume: if(R) goto Start ; else if (A_active | B_active) { if (A_active & A) { A_active=0 ; } if (B_active & B) { B_active=0 ; } if (!(A_active | B_active)) { emit O ; } } goto Out; Start: A_active=1 ; B_active=1 ; Out: loop [ await A || await B ]; emit O; halt every R

15. Old translation schemes Semantically complete Efficient code Intermediate model Explicit FSM Circuits ? Expensive, slow General Cheap, fast General* Cheap, fast Only “acyclic” programs Compiling method Bisimulation (fc2tools) RTL optimizations (SIS) Classical control-flow optimizations Optimization Generated code (without optim.) Very large, Very fast Small Slow Very small Very fast Semantics (acyclic=?) Less powerful optim. Problems Do not scale up well *=sccausal or slow simulation

16. What I wanted • Generate efficient code for “good” programs • Generate code for all programs • Understand cyclicity at a higher level • Inexpensive optimizations based on static analysis • Formalize the efficient approach • New intermediate format/model (GRC) • Hierarchical state representation • Control-flow graph • No specific encoding How

17. GRC format – a small example 1 boot: loop [ await A;emit B || await B ]; emit O; halt every R 0 # 4 await A 3 || await B 2 5 # loop-every halt 6 enter4 exit 1 enter2 enter3 enter5 [1] exit2 0 Inactive[4] [2] R 4 exit4 emit B K0[4] exit3 enter6 K0 emit O A K1[4] K1 Inactive[5] [3] 5 exit5 K0[5] 2 B [6] K1[5]

18. GRC format – utility • Supports the full Esterel semantics • Analysis and optimization at GRC code level • Supports optimized encoding and scheduling algorithms

19. Conclusion • A very brief introduction to Esterel, its compilation, and my work • During my PhD: • New Esterel semantics • Intermediate model for Esterel programs • Static analysis and optimization at this level • General code generation scheme • The best compiler yet

20. Past • PhD at Ecole des Mines de Paris, with Gérard Berry • Complete the work started during the PhD • Investigate aspects related to code distribution (Benveniste et al.) Future

21. Results • Prototype compiler (acyclic case) • Examples: • Turbo channel bus • Berry’s wristwatch • Video generator • Shock absorber • Operating system model • Avionics fuel controller • Avionics cockpit • Man-machine interface Test configuration: PIII/1GHz/128M/Linux gcc-2.96 –O, 1Mcycle random or given