Diego S. Gallo

Active Memory Cube: A processing-in-memory architecture for exascale systems

Many studies point to the difficulty of scaling existing computer architectures to meet the needs of an exascale system (i.e., capable of executing

A crowdsourcing platform for the construction of accessibility maps

10^{18}

CloudBench: Experiment Automation for Cloud Environments

floating-point operations per second), consuming no more than 20 MW in power, by around the year 2020. This paper outlines a new architecture, the Active Memory Cube, which reduces the energy of computation significantly by performing computation in the memory module, rather than moving data through large memory hierarchies to the processor core. The architecture leverages a commercially demonstrated 3D memory stack called the Hybrid Memory Cube, placing sophisticated computational elements on the logic layer below its stack of dynamic random-access memory (DRAM) dies. The paper also describes an Active Memory Cube tuned to the requirements of a scientific exascale system. The computational elements have a vector architecture and are capable of performing a comprehensive set of floating-point and integer instructions, predicated operations, and gather-scatter accesses across memory in the Cube. The paper outlines the software infrastructure used to develop applications and to evaluate the architecture, and describes results of experiments on application kernels, along with performance and power projections.

Taxonomy of Citizen Sensing for Intelligent Urban Infrastructures

We present in this article a crowdsourcing platform that enables the collaborative creation of accessibility maps. The platform provides means for integration of different kind of data, collected automatically or with user intervention, to augment standard maps with accessibility information. The article shows the architecture of the platform, dedicating special attention to the smartphone applications we developed for data collection. The article also describes a preliminar experiment conducted on field, showing how the analysis of data produced by our solution can bring novel insights in accessibility challenges that can be found in cities.

Tracking the Performance Evolution of Blue Gene Systems

The growth in the adoption of cloud computing is driven by distinct and clear benefits for both cloud customers and cloud providers. However, the increase in the number of cloud providers as well as in the variety of offerings from each provider has made it harder for customers to choose. At the same time, the number of options to build a cloud infrastructure, from cloud management platforms to different interconnection and storage technologies, also poses a challenge for cloud providers. In this context, cloud experiments are as necessary as they are labor intensive. Cloud Bench [1] is an open-source framework that automates cloud-scale evaluation and benchmarking through the running of controlled experiments, where complex applications are automatically deployed. Experiments are described through experiment plans, containing directives with enough descriptive power to make the experiment descriptions brief while allowing for customizable multi-parameter variation. Experiments can be executed in multiple clouds using a single interface. Cloud Bench is capable of managing experiments spread across multiple regions and for long periods of time. The modular approach adopted allows it to be easily extended to accommodate new cloud infrastructure APIs and benchmark applications, directly by external users. A built-in data collection system collects, aggregates and stores metrics for cloud management activities (such as VM provisioning and VM image capture) and application runtime information. Experiments can be conducted in a highly controllable fashion, in order to assess the stability, scalability and reliability of multiple cloud configurations. We demonstrate Cloud Benchs main characteristics through the evaluation of an Open Stack installation, including experiments with approximately 1200 simultaneous VMs at an arrival rate of up to 400 VMs/hour.

Smarter board: a community-oriented communication tool

Citizen sensing is a new sensor-based data collection paradigm and is focused on the extraction of data generated by people. Initiatives based on this concept are becoming crucial for designers of intelligent urban infrastructures, since they enable the collection of several types of relevant data that cannot be properly captured by traditional physical sensors. A large number of articles and projects associated with the topic appeared over the last few years, and with them the need for properly classifying and organizing these works. In this paper, we propose a taxonomy of citizen sensing initiatives and illustrate each of its dimensions through a survey of recent articles in the area. The proposed scheme also supports the identification and stimulates the development of projects addressing data collection methodologies that have not been properly explored so far. In addition, we present a platform capable of aggregating, analyzing, and extracting knowledge from data generated by physical and human sensing techniques. Finally, we report a real-world experiment in which we used our platform to map accessibility conditions of streets and sidewalks located in a four square kilometer area in São Paulo, Brazil. Our results show that a full coverage was obtained with the support of eight volunteers after only three hours, hence illustrating the effectiveness of the technology.

Marker-assisted recognition of dynamic content in public spaces

IBM’s Blue Gene supercomputer has evolved through three generations from the original Blue Gene/L to P to Q. A higher level of integration has enabled greater single-core performance, and a larger concurrency per compute node. Although these changes have brought with them a higher overall system peak-performance, no study has examined in detail the evolution of performance across system generations. In this work we make two significant contributions – that of providing a comparative performance analysis across Blue Gene generations using a consistent set of tests, and also in providing a validated performance model of the NEK-Bone proxy application. The combination of empirical analysis and the predictive performance model enable us to not only directly compare measured performance but also allow for a comparison of system configurations that cannot currently be measured. We provide insights into how the changing characteristics of Blue Gene have impacted on the application performance, as well as what future systems may be able to achieve.

Marker-based image recognition of dynamic content for the visually impaired

In this demo we present the Smarter Board, a platform designed to facilitate the creation of a community-focused social network, with a special focus on groups of people with disabilities. The communication is based on text messages, which makes the system easy to use and more accessible to communities where network connections are not well-developed and where the people do not have much experience with more advanced technological tools. The solution also provides a manage interface, through which administrators are able to mediate the messages and the users. We also implemented a matching procedure for the identification of related posts (e.g., it can check if there are compatible car ride offers and car ride requests) in order to make users aware of what is being posted and, even more important, adopt the technology.

Fast full-system execution-driven performance simulator for blue gene/q

In this work we present an image processing-based assistant for helping visually impaired citizens with the task of recognizing dynamic content within fixed layouts of displays in public spaces. Our solution relies on the placement of markers, in order to facilitate the location and recognition of target objects and, at the same time, provide hints to users about how to better position their mobile devices cameras to capture the whole information contained in the display.

A PERFORMANCE ANALYSIS OF THREE GENERATIONS OF BLUE GENE

The access to information displayed in public spaces is a challenge faced by visually impaired people for which image processing techniques have the potential to deliver satisfactory solutions. However, object recognition algorithms must initially locate possible candidates in the images, which is a hard task in complex scenes. In this article, we introduce an image processing technique that relies on the incorporation of markers to panels and boards with fixed layouts displaying dynamic content. The markers allow: a) locating the objects to be recognized; b) correcting perspective in the input images; c) limiting the training set size for supervised learning; and d) guiding the visually impaired by indicating how they should position their devices for adequate pictures. The proposed technique can be used for automatic recognition of texts and images and is suitable for deployment on mobile devices, providing more independence to the citizens. Results of preliminary tests on vending machines show that this method is robust enough to be used in practice.