1 / 19

Scientific Computing

Scientific Computing. By: Fatima Hallak To: Dr. Guy Tel-Zur. SPIN Spin home page. Spin targets efficient software verification. Supports a high level language called PROMELA (a PROcess MEta LAnguage). ً Written in ANSI standard C

maylin
Télécharger la présentation

Scientific Computing

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. Scientific Computing By: Fatima Hallak To: Dr. Guy Tel-Zur

  2. SPIN Spin home page • Spin targets efficient software verification. • Supports a high level language called PROMELA (a PROcess MEta LAnguage). • ًWritten in ANSI standard C • Portable across all versions of Unix, Linux, cygwin, Plan9, Inferno, Solaris, Mac, and Windows

  3. SPIN • V 4 provides direct support for the use of embedded C code as part of model specifications. To directly verify implementation level software specifications, as a driver and as a logic engine to verify high level temporal properties. • V 5 provides direct support for the use of multi-core computers for model checking runs -- supporting both safety and liveness verifications. • Driver for swarm verification (a new form of swarm computing), which can make optimal use of large numbers of available compute cores to leverage parallelism and search diversification techniques.

  4. SPIN’s Functionality • Used to trace logical design errors in distributed systems design, such as operating systems, data communications protocols, switching systems, concurrent algorithms, railway signaling protocols, etc. • Reports on deadlocks, unspecified receptions, flags incompleteness, race conditions, and unwarranted assumptions about the relative speeds of processes

  5. SPIN Supports • Dynamically growing and shrinking numbers of processes, using a rubber state vector technique. • Both rendezvous and buffered message passing, and communication through shared memory. • Mixed systems, using both synchronous and asynchronous communications. • Message channel identifiers for both rendezvous and buffered channels, can be passed from one process to another in messages.

  6. Spin can be used in four main modes • Simulator, allowing for rapid prototyping with a random, guided, or interactive simulations • Exhaustive verifier, capable of rigorously proving the validity of user specified correctness requirements (using partial order reduction theory to optimize the search) either depth-first or breadth-first search. • Proof approximation system that can validate even very large system models with maximal coverage of the state space. • a driver for swarm verification which can make optimal use of large numbers of available compute cores to leverage parallelism and search diversification techniques, which increases the chance of locating defects in very large verification models.

  7. Swarm 3.1 – Verification Script generator for SPIN Syntax: swarm [config_file] [option] * Script that performs many small verification jobs in parallel that can increase the problem coverage for very large verification problems by about an order of magnitude compared to standard bitstate verification runs. It is meant to be used on models for which standard verification with exhaustive, bitstate, hash-compaction etc. either runs out of memory, or takes more time than is available (e.g., days or weeks).

  8. swarm3.1.tar • Swarm uses parallelism and search diversification to reach its objectives.The user can use a configuration file to define: • how many processing cores are available • how much memory can be used • how much time is maximally available • other optional parameter settings. • Based on this information, swarm generates the script that runs as many independent jobs as possible in parallel, without exceeding any of the user-defined constraints. • Swarm can run jobs using local CPU cores or remote machines in a grid network. • Swarm 3.1 is a reimplementation of the algorithm, making more extensive use of search randomization techniques, and taking advantage of some new features implemented for this purpose in Spin Version 5.2 and up

  9. Nondeterministic Finite Automata (NDFA) problem • Nondeterminism is a central concept in CS that appears frequently in applications: grammars of programming languages, algorithms and the interleaving model of concurrency. • The SPIN semantics is existential (a string is accepted if there exists a computation that terminates in an accepting state after reading the entire string) A primer on Model checking by Moti Ben-Ari http://stwww.weizmann.ac.il/g-cs/benari/jspin/

  10. NDFA program • Programmed using the guarded if-command for nondeterministic trasitions. For example: {(q5, a, q7), (q5, a, q3), (q5, b, q5)}. q5: if :: input == ‘a’ -> input = next-symbol; goto q7 :: input == ‘a’ -> input = next-symbol; goto q3 :: input == ‘b’ -> input = next-symbol; goto q5 fi Next, add the alternative: :: end-of-input -> assert(false)

  11. SPIN modes in NDFA • Random simulation is the execution of the NDFA with arbitrary resolution of nondeterministic transitions. • Interactive simulation is the execution of an NDFA with an oracle ensuring that an accepting computation is found • Verification represents the metalevel determination if there exists an accepting computation or not.

  12. Other examples

  13. The Semantics of LTL

  14. bool wantP = false, wantQ= false; Active proctype P() { do :: wantP = true; do :: wantQ -> wantP = false; wantP = true ::else -> break od; wantP = false od } Active proctype Q() { do :: wantQ = true; do :: wantP -> wantQ = false; wantQ = true ::else -> break od; wantQ = false od } The Semantics of LTL

More Related