Navegación

Búsqueda

Búsqueda avanzada

El autor Julio Mariño ha publicado 4 artículo(s):

1 - A Haskell Implementation of a Rule-Based Program Transformation for C Programs

Obtaining good performance when programming heterogeneous computing platforms poses significant challenges for the programmer. We present a program transformation environment, implemented in Haskell, where architecture-agnostic scientific C code is transformed into a functionally equivalent one better suited for a given platform. The transformation rules are formalized in a domain-specific language (STML) that takes care of the syntactic and semantic conditions required to apply a given transformation. STML rules are compiled into Haskell function definitions that operate at AST level. Program properties, to be matched with rule conditions, can be automatically inferred or, alternatively, stated as annotations in the source code. Early experimental results are described.

Autores: Salvador Tamarit / Guillermo Vigueras / Manuel Carro / Julio Mariño / 
Palabras Clave:

2 - Towards Automatic Learning of Heuristics for Mechanical Transformations of Procedural Code

The current trends in next-generation exascale systems go towards integrating a wide range of specialized (co-)processors into traditional supercomputers. Due to the efficiency of heterogeneous systems in terms of Watts and FLOPS per surface unit, opening the access of heterogeneous platforms to a range of users as wide as possible is an important problem to be tackled. However, the integration of heterogeneous, specialized devices increases programming complexity, restricting it to a few experts, and makes porting applications onto different computational infrastructures extremely costly. In order to bridge the gap between the programming needs of heterogeneous systems and the expertise of programmers, program transformation has been proposed elsewhere as a means to ease program generation and adaptation. This brings about several issues such as how to plan a transformation strategy which eventually generates code with increased performance. In this paper we propose a machine learning-based approach to learn heuristics for defining transformation strategies of a program transformation system. Our approach proposes a novel combination of reinforcement learning and classification methods to efficiently tackle the problems inherent to this type of systems. Preliminary results demonstrate the suitability of this approach.

Autores: Guillermo Vigueras  /  Manuel Carro / Salvador Tamarit  / Julio Mariño / 
Palabras Clave:

3 - Tool Demonstration: Testing JSON Web Services Using Jsongen (Demostración)

This article describes a tool, jsongen, which permits testing behavioural aspects of Web Services that communicate using the JSON data format. Provided a characterisation of the JSON data as a JSON schema, the jsongen tool will:(i) automatically derive a QuickCheck (the property-based testing tool)generator which can generate an infinite number of JSON values that validate against the schema, and (ii) provides a generic QuickCheck state machine which is capable of following the (hyper)links documented in the JSON schema, to automatically explore the web service. The default behaviour of the state machine can be easily customized to include web service specific checks. The approach is demonstrated by applying it to the task of testing a simplified web service for banking.

Autores: Ignacio Ballesteros / Luis Eduardo Bueso de Barrio / Lars-Ake Fredlund / Julio Mariño / 
Palabras Clave: Analysis tools - JSON - Testing - web services

4 -