1 / 17

HWU Progress

HWU Progress. Robert Pointon & Greg Michaelson EmBounded Project Meeting Islay. Overview. Compiler Hume problem solving environment Hierarchical Hume Hardware - robots, m32c Related work Future work. Compiler. Unboxing Basic 1+1 for simple types - done Basic wire/heap copy - done

kathie
Télécharger la présentation

HWU Progress

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript


  1. HWU Progress Robert Pointon & Greg Michaelson EmBounded Project Meeting Islay

  2. Overview • Compiler • Hume problem solving environment • Hierarchical Hume • Hardware - robots, m32c • Related work • Future work

  3. Compiler • Unboxing • Basic 1+1 for simple types - done • Basic wire/heap copy - done • I/O - work in progress • ccall - not started (but will follow I/O design) • Structured types on wires, generic operators (show, ==), and exceptions - planning

  4. Hume IDE

  5. Hume IDE • Import/export of Hume programs nearing completion • Priorities now shifted back to compiler • Require efficient Hume performance • especially for vision algorithms • to give realistic real-time demos • Time costing requires better code gen • Currently too generic and unbounded

  6. Hierarchical Hume • Idea: • A box can itself be another Hume program • Need to be able to define termination of nested programs • Can use the existing box patterns to specify pre/post conditions of nested program and thus constraints on behavior! • Observation: • A box need not be a Hume program, could be anything in any language, thus gives component based design!

  7. Features • Specify termination condition for the scheduling child boxes, thus: • Can reason about parent box without considering it’s children • Can reason about child boxes without considering parent box context • Namespace for functions and types • Stateless boxes means no I/O defined in nested boxes - still being debated… • Unhandled exceptions propagate up hierarchy

  8. Example box foo in (i::int 32) out (o::string, p::int 32) match (i) -> (_, i+1) -- ‘_’ means some result, so RHS is a pattern!? boxes -- start nesting box bar in (i::int 32) out (o::string, p::int 32) match (x) -> ((x, ’\n’) as string, x+1); wire bar (i) (o, p); -- parent box name isn’t required end; -- finish nesting

  9. Status • Several papers written with Gudmund Grov • Supported in Hume front-end - parser, macro expansion, etc • Working in Hume interpreter! • Not supported in Hume compiler • would need work on type checker (function/type namespace) and HAM code generation • Would require re-think of C code generation • Suggest delaying until any later redesign for multi-threading and better box scheduling… not this project…

  10. Hardware - robot • Algorithms are a bit too slow for convincing demos • Unboxing will help • Issues with I/O • Spins rather than waits, so multiple Hume programs tend to use up CPU power quickly • Everything via sockets • Must remember to turn off all debugging, and allow the C compiler to optimize!

  11. Hardware - m32c • Expansion Memory • £@&%^$ tech support from companies… • Two weeks time from electronics RA and no further forward… suggestions? • I/O and Debugging • New ModSDK board is much nicer! • LCD, extra serial ports, etc • But m32c87 - how different from ‘85’?

  12. Related PhD Work • Gudmund Grov • Hume coordination layer in TLA • scheduling proof in TLA • self-out & Hierarchical Hume are conservative extensions of super-step scheduling • TLA in Isabelle • box calculus • TFP 8 paper accepted

  13. Related PhD Work • Chunxu Liu • HW-Hume in Isabelle • correctness proofs of multi-box H-W Hume programs from standard repertoire • writing PhD

  14. SEAS DTC Project • Hume WCET analysis/evaluation • object tracking algorithms in Hume & Matlab/C++ • Hume integrated in Pioneer AT-P3 system

  15. SEAS DTC Project C motor API Hume Java GUI WiFi WiFi C image API Hume C pan/tilt API Hume Laptop Pioneer AT-P3

  16. Future work • complete unboxing • complete base IDE • deliverables • Robert Pointon leaves Dec 2007 • SEAS DTC ends Dec 2007

  17. Future Work • EU FP7 EmBounded 2... • EPSRC – 2007/10 • sensor networks/DSP/dynamic configuration • HWU & St A with Edinburgh & Queens Uni Belfast • SEAS DTC – 2008/11 • compositional framework/component based simulation • HWU & StA with Waterfall Solutions

More Related