Executable engines for relational model-transformation languages evolve continuously because of language extension, performance improvement and bug fixes. While new versions generally change the engine semantics, end-...
详细信息
Executable engines for relational model-transformation languages evolve continuously because of language extension, performance improvement and bug fixes. While new versions generally change the engine semantics, end-users expect to get backward-compatibility guarantees, so that existing transformations do not need to be adapted at every engine update. The CoqTL model-transformation language allows users to define model transformations, theorems on their behavior and machine-checked proofs of these theorems in Coq. Backward-compatibility for CoqTL involves also the preservation of these proofs. However, proof preservation is challenging, as proofs are easily broken even by small refactorings of the code they verify. In this paper, we present the solution we designed for the evolution of CoqTL. We provide a deep specification of the transformation engine, including a set of theorems that must hold against the engine implementation. Then, at each milestone in the engine development, we certify the new version of the engine against this specification, by providing proofs of the impacted theorems. The certification formally guarantees end-users that all the proofs they write using the provided theorems will be preserved through engine updates. We illustrate the structure of the deep specification theorems, we produce a machine-checked certification of three versions of CoqTL against it, and we show examples of user proofs that leverage this specification and are thus preserved through the updates. Finally, we discuss the evolution of the deep specification by an extension mechanism, we present an evolution that introduces trace links in the specification, and we show which user proofs are preserved through specification evolutions.
Constraint-logic object-oriented programming facilitates the integrated development of business software that occasionally solves constraint-logic problems or makes other use of structured search. So far, work in cons...
详细信息
ISBN:
(纸本)9781450386890
Constraint-logic object-oriented programming facilitates the integrated development of business software that occasionally solves constraint-logic problems or makes other use of structured search. So far, work in constraint-logic object-oriented programming has been limited to considering constraints that only involve logic variables of primitive types;in particular, boolean, integer, and floating-point numbers. However, the availability of object-oriented features calls for the option to use logic variables in lieu of objects as well. Therefore, support for reference-type logic variables (or free objects) is required. With the work at hand, we add support for free objects to a Java-based constraint-logic object-oriented language, Muli. Allowing free objects in statements and expressions results in novel interactions with objects at run time, for instance, non-deterministic execution of polymorphic method invocations (taking arbitrary class hierarchies into account) and non-deterministic type operations. At run time, this is supported by a dynamic type constraint that restricts the possible types of a free object at non-deterministic choices.
Executable engines for relational model-transformation languages evolve continuously because of language extension, performance improvement and bug fixes. While new versions generally change the engine semantics, end-...
详细信息
ISBN:
(纸本)9781450370196
Executable engines for relational model-transformation languages evolve continuously because of language extension, performance improvement and bug fixes. While new versions generally change the engine semantics, end-users expect to get backward-compatibility guarantees, so that existing transformations do not need to be adapted at every engine update. The CoqTL model-transformation language allows users to define model transformations, theorems on their behavior and machinechecked proofs of these theorems in Coq. Backward-compatibility for CoqTL involves also the preservation of these proofs. However, proof preservation is challenging, as proofs are easily broken even by small refactorings of the code they verify. In this paper we present the solution we designed for the evolution of CoqTL, and by extension, of rule-based transformation engines. We provide a deep specification of the transformation engine, including a set of theorems that must hold against the engine implementation. Then, at each milestone in the engine development, we certify the new version of the engine against this specification, by providing proofs of the impacted theorems. The certification formally guarantees end-users that all the proofs they write using the provided theorems will be preserved through engine updates. We illustrate the structure of the deep specification theorems, we produce a machine-checked certification of three versions of CoqTL against it, and we show examples of user theorems that leverage this specification and are thus preserved through the updates.
Modern accelerator programming frameworks, such as OpenCL (TM), organise threads into work-groups. Remote-scope promotion (RSP) is a language extension recently proposed by AMD researchers that is designed to enable a...
详细信息
ISBN:
(纸本)9781450336895
Modern accelerator programming frameworks, such as OpenCL (TM), organise threads into work-groups. Remote-scope promotion (RSP) is a language extension recently proposed by AMD researchers that is designed to enable applications, for the first time, both to optimise for the common case of intra-work-group communication (using memory scopes to provide consistency only within a work-group) and to allow occasional inter-work-group communication (as required, for instance, to support the popular load-balancing idiom of work stealing). We present the first formal, axiomatic memory model of OpenCL extended with RSP. We have extended the HERD memory model simulator with support for OpenCL kernels that exploit RSP, and used it to discover bugs in several litmus tests and a work-stealing queue, that have been used previously in the study of RSP. We have also formalised the proposed GPU implementation of RSP. The formalisation process allowed us to identify bugs in the description of RSP that could result in well-synchronised programs experiencing memory inconsistencies. We present and prove sound a new implementation of RSP that incorporates bug fixes and requires less non-standard hardware than the original implementation. This work, a collaboration between academia and industry, clearly demonstrates how, when designing hardware support for a new concurrent language feature, the early application of formal tools and techniques can help to prevent errors, such as those we have found, from making it into silicon.
PIC microcontrollers are low-cost programmable integrated circuits, consume very little energy, but are hard to program due to very little available resources. They are traditionally programmed using low-level languag...
详细信息
ISBN:
(数字)9783319196862
ISBN:
(纸本)9783319196862;9783319196855
PIC microcontrollers are low-cost programmable integrated circuits, consume very little energy, but are hard to program due to very little available resources. They are traditionally programmed using low-level languages (e.g., assembler or subsets of C), which provide very few safeguards if any. This paper presents the issues we had to solve to successfully port a modern multi-paradigm general-purpose programminglanguage, which notably provides automatic memory management and strong static type checking, to this rather peculiar hardware.
Modern accelerator programming frameworks, such as OpenCL (TM), organise threads into work-groups. Remote-scope promotion (RSP) is a language extension recently proposed by AMD researchers that is designed to enable a...
详细信息
Modern accelerator programming frameworks, such as OpenCL (TM), organise threads into work-groups. Remote-scope promotion (RSP) is a language extension recently proposed by AMD researchers that is designed to enable applications, for the first time, both to optimise for the common case of intra-work-group communication (using memory scopes to provide consistency only within a work-group) and to allow occasional inter-work-group communication (as required, for instance, to support the popular load-balancing idiom of work stealing). We present the first formal, axiomatic memory model of OpenCL extended with RSP. We have extended the HERD memory model simulator with support for OpenCL kernels that exploit RSP, and used it to discover bugs in several litmus tests and a work-stealing queue, that have been used previously in the study of RSP. We have also formalised the proposed GPU implementation of RSP. The formalisation process allowed us to identify bugs in the description of RSP that could result in well-synchronised programs experiencing memory inconsistencies. We present and prove sound a new implementation of RSP that incorporates bug fixes and requires less non-standard hardware than the original implementation. This work, a collaboration between academia and industry, clearly demonstrates how, when designing hardware support for a new concurrent language feature, the early application of formal tools and techniques can help to prevent errors, such as those we have found, from making it into silicon.
This paper presents the integration of the optimization known as dynamic cut within the functional-logic system TOY. The implementation automatically detects deterministic functions at compile time, and includes in th...
详细信息
This paper presents the integration of the optimization known as dynamic cut within the functional-logic system TOY. The implementation automatically detects deterministic functions at compile time, and includes in the generated code the test for detecting at run-time the computations that can actually be pruned. The outcome is a much better performance when executing deterministic functions including either or-branches in their definitional trees or extra variables in their conditions, with no serious overhead in the rest of the computations. The paper also proves the correctness of the criterion used for detecting deterministic functions w.r.t. the semantic calculus CRWL.
This paper develops a formalism that precisely characterizes when class tables are required for C++ memory layouts. A memory layout is a particular choice of data structures for implementing run-time support for objec...
详细信息
This paper develops a formalism that precisely characterizes when class tables are required for C++ memory layouts. A memory layout is a particular choice of data structures for implementing run-time support for object-oriented languages. We use this formalism to quantify and evaluate, on a set of benchmarks, the space overhead for a set of C++ memory layouts. In particular, this paper studies the space overhead due to three language features: virtual dispatch, virtual inheritance, and dynamic typing. To date, there has been no scientific quantification or evaluation of C++ memory layouts. Our approach can help C++ implementors. This work has already influenced the memory layout design choices in IBM's Visual Age C++ V5 compiler. Applying our approach to a set of five benchmarks, we demonstrate that the impact of object-oriented space overhead can vary dramatically between applications (ranging from 0.42% to 99.79% for our benchmarks). In particular, applications whose object space is dominated by instances of classes that heavily use object-oriented language features will be significantly impacted by the choice of a memory layout. Copyright (C) 2003 John Wiley Sons, Ltd.
Previous attempts at garbage collection in uncooperative environments have generally used conservative or mostly-conservative approaches. We describe a technique for doing fully type-accurate garbage collection in an ...
详细信息
Previous attempts at garbage collection in uncooperative environments have generally used conservative or mostly-conservative approaches. We describe a technique for doing fully type-accurate garbage collection in an uncooperative environment, using a "shadow stack" to link structs of pointer-containing variables, together with the data or code needed to trace them. We have implemented this in the Mercury compiler, which generates C code, and present preliminary performance data on the overheads of this technique. We also show how this technique can be extended to handle multithreaded applications.
Many logic programmingimplementations compile to C, but they compile to very low-level C, and thus discard many of the advantages of compiling to a high-level language. We describe an alternative approach to compilin...
详细信息
ISBN:
(纸本)3540433694
Many logic programmingimplementations compile to C, but they compile to very low-level C, and thus discard many of the advantages of compiling to a high-level language. We describe an alternative approach to compiling logic programs to C, based on continuation passing, that we have used in a new back-end for the Mercury compiler. The new approach compiles to much higher-level C code, which means the compiler back-end and run-time system can be considerably simpler. We present a formal schema for the transformation, and give benchmark results which show that this approach delivers performance that is more than competitive with the fastest previous implementation, with greater simplicity and better portability and interoperability. The approach we describe can also be used for compiling to other target languages, such as IL (the Microsoft NET intermediate language).
暂无评论