Dataset Information

Tool support for software lookup table optimization.

ABSTRACT: A number of scientific applications are performance-limited by expressions that repeatedly call costly elementary functions. Lookup table (LUT) optimization accelerates the evaluation of such functions by reusing previously computed results. LUT methods can speed up applications that tolerate an approximation of function results, thereby achieving a high level of fuzzy reuse. One problem with LUT optimization is the difficulty of controlling the tradeoff between performance and accuracy. The current practice of manual LUT optimization adds programming effort by requiring extensive experimentation to make this tradeoff, and such hand tuning can obfuscate algorithms. In this paper we describe a methodology and tool implementation to improve the application of software LUT optimization. Our Mesa tool implements source-to-source transformations for C or C++ code to automate the tedious and error-prone aspects of LUT generation such as domain profiling, error analysis, and code generation. We evaluate Mesa with five scientific applications. Our results show a performance improvement of 3.0 × and 6.9 × for two molecular biology algorithms, 1.4 × for a molecular dynamics program, 2.1 × to 2.8 × for a neural network application, and 4.6 × for a hydrology calculation. We find that Mesa enables LUT optimization with more control over accuracy and less effort than manual approaches.

SUBMITTER: Wilcox C

PROVIDER: S-EPMC3922221 | biostudies-literature | 2011 Dec

REPOSITORIES: biostudies-literature

ACCESS DATA

Publications

Tool support for software lookup table optimization.

Wilcox Chris C Strout Michelle Mills MM Bieman James M JM

Scientific programming 20111201 4

A number of scientific applications are performance-limited by expressions that repeatedly call costly elementary functions. Lookup table (LUT) optimization accelerates the evaluation of such functions by reusing previously computed results. LUT methods can speed up applications that tolerate an approximation of function results, thereby achieving a high level of <i>fuzzy reuse</i>. One problem with LUT optimization is the difficulty of controlling the tradeoff between performance and accuracy. ...[more]

PMID: 24532963

Similar Datasets

Project description:BackgroundSince its inception, proteomics has essentially operated in a discovery mode with the goal of identifying and quantifying the maximal number of proteins in a sample. Increasingly, proteomic measurements are also supporting hypothesis-driven studies, in which a predetermined set of proteins is consistently detected and quantified in multiple samples. Selected reaction monitoring (SRM) is a targeted mass spectrometric technique that supports the detection and quantification of specific proteins in complex samples at high sensitivity and reproducibility. Here, we describe ATAQS, an integrated software platform that supports all stages of targeted, SRM-based proteomics experiments including target selection, transition optimization and post acquisition data analysis. This software will significantly facilitate the use of targeted proteomic techniques and contribute to the generation of highly sensitive, reproducible and complete datasets that are particularly critical for the discovery and validation of targets in hypothesis-driven studies in systems biology.ResultWe introduce a new open source software pipeline, ATAQS (Automated and Targeted Analysis with Quantitative SRM), which consists of a number of modules that collectively support the SRM assay development workflow for targeted proteomic experiments (project management and generation of protein, peptide and transitions and the validation of peptide detection by SRM). ATAQS provides a flexible pipeline for end-users by allowing the workflow to start or end at any point of the pipeline, and for computational biologists, by enabling the easy extension of java algorithm classes for their own algorithm plug-in or connection via an external web site.This integrated system supports all steps in a SRM-based experiment and provides a user-friendly GUI that can be run by any operating system that allows the installation of the Mozilla Firefox web browser.ConclusionsTargeted proteomics via SRM is a powerful new technique that enables the reproducible and accurate identification and quantification of sets of proteins of interest. ATAQS is the first open-source software that supports all steps of the targeted proteomics workflow. ATAQS also provides software API (Application Program Interface) documentation that enables the addition of new algorithms to each of the workflow steps. The software, installation guide and sample dataset can be found in http://tools.proteomecenter.org/ATAQS/ATAQS.html.

Project description:BACKGROUND:Few studies have analyzed gains in using computers in speech and language therapy interventions for children with speech and/or language disorders when compared to a control group, but virtual tutors and computer-based visual feedback have been gaining interest in the literature. Previous systematic reviews mainly focused on development technological details of computer-based speech training systems or the potential of integrating mobile technology into education and rehabilitation, but recent systematic reviews have also evaluated the efficacy of computer-based speech and language therapy for children and how digital technology can support different activities, at school or elsewhere. OBJECTIVE:This study aimed to analyze a continuous communication and joint team approach to develop solutions focused on the real needs of end users, which digitally emulate reliable and validated physical intervention materials for children with speech sound disorders (SSD). METHODS:The Table to Tablet (T2T) software was developed using a design-based research methodology, which included four phases: activities development; ethnographic pretesting with a sample from the target population; software development; and beta-testing. The technology used to develop the software, the method used to ensure satisfaction and replay ability of the intervention materials, and results from the ethnographic and beta-testing phases are presented. RESULTS:Nineteen activities were developed during the first phase, which were then tested, with 7 service users, using a physical prototype. The beta-test approach included extensive testing and reformulation, supported by direct, nonparticipant observation and data collection using a questionnaire designed for children. Feedback was used to improve the software and interaction with users. CONCLUSIONS:The use of T2T-based intervention programmes by speech and language therapists (SLTs) will allow these professionals to make a better and more effective communication intervention, based on proven methodologies, that coexists in a structured physical and a digital version. These versions provide a full, 6-week intervention program, with minimal effort in preparing the session by the SLTs while delivering a very consistent intervention, with high replay value. A continuous communication and joint team approach was beneficial to the project and to the development of a solution focused on the real needs of SLTs and children with SSD. All problems were approached as a team with different skills and expertise, which minimized errors (eg, the developer making choices that would save him from spending time doing something that would not be used) and time spent. To add to this, the importance of integrating the end users as testers and collecting their opinions and actions per session allowed the production of better-targeted activities. TRIAL REGISTRATION:ClinicalTrials.gov NCT02490826; https://clinicaltrials.gov/ct2/show/NCT02490826.

Dataset Information

Tool support for software lookup table optimization.

Publications

Tool support for software lookup table optimization.

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure

Tweets