Project

General

Profile

Development Milestones

Version 2 (deadline: 31/08/2014)

MicroTESK v2 should fully support instruction set specifications and random/combinatorial template-based test program generation. This is the minimal functionality that makes the MicroTESK tool useful.

Version 2.0 30/04/2014 Features: nML-based instruction set specification, test situations, exceptions, initialization code, branching instructions
  • Test Situations
    • Manual description of test situations in Java and XML should be supported
    • Test situations should be accessible from test templates (corresponding solvers should be invoked, and generated data should be substituted into the test program)
    • Each instruction should have a top-level test situation (by default, the top-level situation is Random)
    • There should be a configuration file assigning top-level test situations to instructions (Random, Zero, etc.)
  • Initialization Code
    • Manual specification of initialization code (for each access mode) should be supported
    • Automated extraction of initialization code (for each access mode) should be implemented
  • Exceptions
    • Mechanism for specifying exceptions should be clarified (probably, a predefined function, like RaiseException, should be introduced)
    • Means for describing exception handling logic should be added
  • Branch Instructions
    • Facilities for defining/using labels in test templates should be revisited
    • Combinatorial test program generation based on test templates with branch instructions should be implemented
  • Examples
    • A test template with branch instructions (combinatorial generation)
  • Documentation
    • All new facilities should be documented
Version 2.1 31/05/2014 Features: VLIW specification and testing, floating-point arithmetic, register aliases, self-checking test program generation
  • Specification
    • Means for specifying VLIW architectures (e.g., Elbrus) should be added
    • Floating-point arithmetic should be supported
    • Auxiliary nML constructs, like alias, should be enabled
    • Mechanisms for calling operations in operations should be implemented
  • Testing
    • Mechanisms for calling test templates in test templates should be implemented
    • Generation of self-checking test programs should be supported
    • Means for writting test templates for VLIW architectures should be added
  • Examples
    • A fragment of the Elbrus architecture should specified
    • Some test templates for the Elbrus architecture should be written
  • Documentation
    • All new facilities should be documented
Version 2.2 30/06/2014 Features: test situation composition/decomposition/iteration, test coverage extraction, constrained-random test data generation
  • Test Situation Composition
    • Disjunctive composition of test situations should be supported (random choice of a test situation based on biases)
    • Conjunctive composition of test situations should be supported (there should possibility to specify hard and soft constraints)
  • Test Situation Decomposition
    • Means for splitting test situations into disjunctions of implicants should be implemented (BDD-based and other test generation algorithms can be applied): rule(situation)
    • The set of decomposition rules should be extensible
  • Test Situation Iteration
    • Means for iterating test situations should be implemented: iterate(disjuntive_situation), iterator(rule(situation)) or iterate(situation1, situation2, ...)
  • Test Coverage Extraction
    • For each instruction and for each path in its CFG, the corresponding test situation should be constructed
    • For each instruction, the top-level test situation should be constructed as the disjunctive composition of the extracted test situations
  • Examples
    • A test situation composition example should be added
    • A test situation decomposition example should be added
    • A test situation iteration example should be added
  • Documentation
    • All new facilities should be documented
Version 2.3 31/08/2014 Features: integration with TestBase, Eclipse-based development environment, test program tracing
  • TestBase Project
    • Functions for structuring/composing/decomposing/... test situations should be separated into the TestBase project
    • Additional test generation facilities should be implemented
  • MicroTESK IDE
    • An Eclipse plugin should be implemented (including an nML editor, instructions/situations viewers, test template wizards, compiler/generator launchers, etc.)
  • Tracing
    • A trace format should be developed
    • Trace utilities should be implemented and integrated into the test program generator
  • Documentation
    • All new facilities should be documented
  • Promotion
    • The miniMIPS microprocessor should be specified
    • Test templates for the miniMIPS microprocessor should be written
    • The test program generator should be added into the miniMIPS project at OpenCores

Version 3 (Deadline: 31/08/2015)

MicroTESK v3 should support microarchitectural specifications and automated test template generation. The tool should be extended to complex multicore microprocessors with cache coherence support.

Version 3.0 31/12/2014 Features: memory management, caching, address translation
  • Microarchitectural Specifications
    • Specification of memory management units should supported
Version 3.1 31/05/2015 Features: pipelining, microarchitectural networks
  • Internal Representation
    • Internal representation should be unified with the Retrascope's one
Version 3.2 31/08/2015 Features: multicore microprocessors, Promela, cache coherence
  • Multicore Microprocessors
    • Integration with Promela specifications should be implemented

To be continued...