Optimización de la Entrada Salida mediante librerías y lenguajes paralelos

Uno de los grandes retos de la HPC (High Performance Computing) consiste en optimizar el subsistema de Entrada/Salida, (E/S), o I/O (Input/Output). Ken Batcher resume este hecho en la siguiente frase: "Un supercomputador es un dispositivo que convierte los problemas limitados por la potencia de cálculo en problemas limitados por la E/S" ("A Supercomputer is a device for turning compute-bound problems into I/O-bound problems") . En otras palabras, el cuello de botella ya no reside tanto en el procesamiento de los datos como en la disponibilidad de los mismos. Además, este problema se exacerbará con la llegada del Exascale y la popularización de las aplicaciones Big Data. En este contexto, esta tesis contribuye a mejorar el rendimiento y la facilidad de uso del subsistema de E/S de los sistemas de supercomputación. Principalmente se proponen dos contribuciones al respecto: i) una interfaz de E/S desarrollada para el lenguaje Chapel que mejora la productividad del programador a la hora de codificar las operaciones de E/S; y ii) una implementación optimizada del almacenamiento de datos de secuencias genéticas. Con más detalle, la primera contribución estudia y analiza distintas optimizaciones de la E/S en Chapel, al tiempo que provee a los usuarios de una interfaz simple para el acceso paralelo y distribuido a los datos contenidos en ficheros. Por tanto, contribuimos tanto a aumentar la productividad de los desarrolladores, como a que la implementación sea lo más óptima posible. La segunda contribución también se enmarca dentro de los problemas de E/S, pero en este caso se centra en mejorar el almacenamiento de los datos de secuencias genéticas, incluyendo su compresión, y en permitir un uso eficiente de esos datos por parte de las aplicaciones existentes, permitiendo una recuperación eficiente tanto de forma secuencial como aleatoria. Adicionalmente, proponemos una implementación paralela basada en Chapel.

Multi-sensor Cloud and Aerosol Retrieval Simulator and Its Applications

Executing a cloud or aerosol physical properties retrieval algorithm from controlled synthetic data is an important step in retrieval algorithm development. Synthetic data can help answer questions about the sensitivity and performance of the algorithm or aid in determining how an existing retrieval algorithm may perform with a planned sensor. Synthetic data can also help in solving issues that may have surfaced in the retrieval results. Synthetic data become very important when other validation methods, such as field campaigns,are of limited scope. These tend to be of relatively short duration and often are costly. Ground stations have limited spatial coverage whilesynthetic data can cover large spatial and temporal scales and a wide variety of conditions at a low cost. In this work I develop an advanced cloud and aerosol retrieval simulator for the MODIS instrument, also known as Multi-sensor Cloud and Aerosol Retrieval Simulator (MCARS). In a close collaboration with the modeling community I have seamlessly combined the GEOS-5 global climate model with the DISORT radiative transfer code, widely used by the remote sensing community, with the observations from the MODIS instrument to create the simulator. With the MCARS simulator it was then possible to solve the long standing issue with the MODIS aerosol optical depth retrievals that had a low bias for smoke aerosols. MODIS aerosol retrieval did not account for effects of humidity on smoke aerosols. The MCARS simulator also revealed an issue that has not been recognized previously, namely,the value of fine mode fraction could create a linear dependence between retrieved aerosol optical depth and land surface reflectance. MCARS provided the ability to examine aerosol retrievals against “ground truth” for hundreds of thousands of simultaneous samples for an area covered by only three AERONET ground stations. Findings from MCARS are already being used to improve the performance of operational MODIS aerosol properties retrieval algorithms. The modeling community will use the MCARS data to create new parameterizations for aerosol properties as a function of properties of the atmospheric column and gain the ability to correct any assimilated retrieval data that may display similar dependencies in comparisons with ground measurements.

HPC Accelerators with 3D Memory

After a decade evolving in the High Performance Computing arena, GPU-equipped supercomputers have con- quered the top500 and green500 lists, providing us unprecedented levels of computational power and memory bandwidth. This year, major vendors have introduced new accelerators based on 3D memory, like Xeon Phi Knights Landing by Intel and Pascal architecture by Nvidia. This paper reviews hardware features of those new HPC accelerators and unveils potential performance for scientific applications, with an emphasis on Hybrid Memory Cube (HMC) and High Bandwidth Memory (HBM) used by commercial products according to roadmaps already announced.

POSE ESTIMATION AND 3D RECONSTRUCTION USING SENSOR FUSION

A camera maps 3-dimensional (3D) world space to a 2-dimensional (2D) image space. In the process it loses the depth information, i.e., the distance from the camera focal point to the imaged objects. It is impossible to recover this information from a single image. However, by using two or more images from different viewing angles this information can be recovered, which in turn can be used to obtain the pose (position and orientation) of the camera. Using this pose, a 3D reconstruction of imaged objects in the world can be computed. Numerous algorithms have been proposed and implemented to solve the above problem; these algorithms are commonly called Structure from Motion (SfM). State-of-the-art SfM techniques have been shown to give promising results. However, unlike a Global Positioning System (GPS) or an Inertial Measurement Unit (IMU) which directly give the position and orientation respectively, the camera system estimates it after implementing SfM as mentioned above. This makes the pose obtained from a camera highly sensitive to the images captured and other effects, such as low lighting conditions, poor focus or improper viewing angles. In some applications, for example, an Unmanned Aerial Vehicle (UAV) inspecting a bridge or a robot mapping an environment using Simultaneous Localization and Mapping (SLAM), it is often difficult to capture images with ideal conditions. This report examines the use of SfM methods in such applications and the role of combining multiple sensors, viz., sensor fusion, to achieve more accurate and usable position and reconstruction information. This project investigates the role of sensor fusion in accurately estimating the pose of a camera for the application of 3D reconstruction of a scene. The first set of experiments is conducted in a motion capture room. These results are assumed as ground truth in order to evaluate the strengths and weaknesses of each sensor and to map their coordinate systems. Then a number of scenarios are targeted where SfM fails. The pose estimates obtained from SfM are replaced by those obtained from other sensors and the 3D reconstruction is completed. Quantitative and qualitative comparisons are made between the 3D reconstruction obtained by using only a camera versus that obtained by using the camera along with a LIDAR and/or an IMU. Additionally, the project also works towards the performance issue faced while handling large data sets of high-resolution images by implementing the system on the Superior high performance computing cluster at Michigan Technological University.

Leakage Temperature Dependency Aware Real-Time Scheduling for Power and Thermal Optimization

Catering to society’s demand for high performance computing, billions of transistors are now integrated on IC chips to deliver unprecedented performances. With increasing transistor density, the power consumption/density is growing exponentially. The increasing power consumption directly translates to the high chip temperature, which not only raises the packaging/cooling costs, but also degrades the performance/reliability and life span of the computing systems. Moreover, high chip temperature also greatly increases the leakage power consumption, which is becoming more and more significant with the continuous scaling of the transistor size. As the semiconductor industry continues to evolve, power and thermal challenges have become the most critical challenges in the design of new generations of computing systems. In this dissertation, we addressed the power/thermal issues from the system-level perspective. Specifically, we sought to employ real-time scheduling methods to optimize the power/thermal efficiency of the real-time computing systems, with leakage/ temperature dependency taken into consideration. In our research, we first explored the fundamental principles on how to employ dynamic voltage scaling (DVS) techniques to reduce the peak operating temperature when running a real-time application on a single core platform. We further proposed a novel real-time scheduling method, “M-Oscillations” to reduce the peak temperature when scheduling a hard real-time periodic task set. We also developed three checking methods to guarantee the feasibility of a periodic real-time schedule under peak temperature constraint. We further extended our research from single core platform to multi-core platform. We investigated the energy estimation problem on the multi-core platforms and developed a light weight and accurate method to calculate the energy consumption for a given voltage schedule on a multi-core platform. Finally, we concluded the dissertation with elaborated discussions of future extensions of our research.

Optimization of Pattern Matching Algorithms for Multi- and Many-Core Platforms

Image and video compression play a major role in the world today, allowing the storage and transmission of large multimedia content volumes. However, the processing of this information requires high computational resources, hence the improvement of the computational performance of these compression algorithms is very important. The Multidimensional Multiscale Parser (MMP) is a pattern-matching-based compression algorithm for multimedia contents, namely images, achieving high compression ratios, maintaining good image quality, Rodrigues et al. [2008]. However, in comparison with other existing algorithms, this algorithm takes some time to execute. Therefore, two parallel implementations for GPUs were proposed by Ribeiro [2016] and Silva [2015] in CUDA and OpenCL-GPU, respectively. In this dissertation, to complement the referred work, we propose two parallel versions that run the MMP algorithm in CPU: one resorting to OpenMP and another that converts the existing OpenCL-GPU into OpenCL-CPU. The proposed solutions are able to improve the computational performance of MMP by 3 and 2:7 , respectively. The High Efficiency Video Coding (HEVC/H.265) is the most recent standard for compression of image and video. Its impressive compression performance, makes it a target for many adaptations, particularly for holoscopic image/video processing (or light field). Some of the proposed modifications to encode this new multimedia content are based on geometry-based disparity compensations (SS), developed by Conti et al. [2014], and a Geometric Transformations (GT) module, proposed by Monteiro et al. [2015]. These compression algorithms for holoscopic images based on HEVC present an implementation of specific search for similar micro-images that is more efficient than the one performed by HEVC, but its implementation is considerably slower than HEVC. In order to enable better execution times, we choose to use the OpenCL API as the GPU enabling language in order to increase the module performance. With its most costly setting, we are able to reduce the GT module execution time from 6.9 days to less then 4 hours, effectively attaining a speedup of 45 .

Constraint Programming-based Job Dispatching for Modern HPC Applications

A High-Performance Computing job dispatcher is a critical software that assigns the finite computing resources to submitted jobs. This resource assignment over time is known as the on-line job dispatching problem in HPC systems. The fact the problem is on-line means that solutions must be computed in real-time, and their required time cannot exceed some threshold to do not affect the normal system functioning. In addition, a job dispatcher must deal with a lot of uncertainty: submission times, the number of requested resources, and duration of jobs. Heuristic-based techniques have been broadly used in HPC systems, at the cost of achieving (sub-)optimal solutions in a short time. However, the scheduling and resource allocation components are separated, thus generates a decoupled decision that may cause a performance loss. Optimization-based techniques are less used for this problem, although they can significantly improve the performance of HPC systems at the expense of higher computation time. Nowadays, HPC systems are being used for modern applications, such as big data analytics and predictive model building, that employ, in general, many short jobs. However, this information is unknown at dispatching time, and job dispatchers need to process large numbers of them quickly while ensuring high Quality-of-Service (QoS) levels. Constraint Programming (CP) has been shown to be an effective approach to tackle job dispatching problems. However, state-of-the-art CP-based job dispatchers are unable to satisfy the challenges of on-line dispatching, such as generate dispatching decisions in a brief period and integrate current and past information of the housing system. Given the previous reasons, we propose CP-based dispatchers that are more suitable for HPC systems running modern applications, generating on-line dispatching decisions in a proper time and are able to make effective use of job duration predictions to improve QoS levels, especially for workloads dominated by short jobs.

Laying the path to consumer-level immersive simulation environment

Since the end of the long winter of virtual reality (VR) at the beginning of the 2010 decade, many improvements have been made in terms of hardware technologies and software platforms performances and costs. Many expect such trend will continue, pushing the penetration rate of virtual reality headsets to skyrocket at some point in the future, just as mobile platforms did before. In the meantime, virtual reality is slowly transitioning from a specialized laboratory-only technology, to a consumer electronics appliance, opening interesting opportunities and challenges. In this transition, two interesting research questions amount to how 2D-based content and applications may benefit (or be hurt) by the adoption of 3D-based immersive environments and to how to proficiently support such integration. Acknowledging the relevance of the former, we here consider the latter question, focusing our attention on the diversified family of PC-based simulation tools and platforms. VR-based visualization is, in fact, widely understood and appreciated in the simulation arena, but mainly confined to high performance computing laboratories. Our contribution here aims at characterizing the simulation tools which could benefit from immersive interfaces, along with a general framework and a preliminary implementation which may be put to good use to support their transition from uniquely 2D to blended 2D/3D environments.

The European legal approach to Open Science and research data

This dissertation proposes an analysis of the governance of the European scientific research, focusing on the emergence of the Open Science paradigm: a new way of doing science, oriented towards the openness of every phase of the scientific research process, able to take full advantage of the digital ICTs. The emergence of this paradigm is relatively recent, but in the last years it has become increasingly relevant. The European institutions expressed a clear intention to embrace the Open Science paradigm (eg., think about the European Open Science Cloud, EOSC; or the establishment of the Horizon Europe programme). This dissertation provides a conceptual framework for the multiple interventions of the European institutions in the field of Open Science, addressing the major legal challenges of its implementation. The study investigates the notion of Open Science, proposing a definition that takes into account all its dimensions related to the human and fundamental rights framework in which Open Science is grounded. The inquiry addresses the legal challenges related to the openness of research data, in light of the European Open Data framework and the impact of the GDPR on the context of Open Science. The last part of the study is devoted to the infrastructural dimension of the Open Science paradigm, exploring the e-infrastructures. The focus is on a specific type of computational infrastructure: the High Performance Computing (HPC) facility. The adoption of HPC for research is analysed from the European perspective, investigating the EuroHPC project, and the local perspective, proposing the case study of the HPC facility of the University of Luxembourg, the ULHPC. This dissertation intends to underline the relevance of the legal coordination approach, between all actors and phases of the process, in order to develop and implement the Open Science paradigm, adhering to the underlying human and fundamental rights.

Deployment of a data analysis workflow of the ATLAS experiment on HPC systems

LHC experiments produce an enormous amount of data, estimated of the order of a few PetaBytes per year. Data management takes place using the Worldwide LHC Computing Grid (WLCG) grid infrastructure, both for storage and processing operations. However, in recent years, many more resources are available on High Performance Computing (HPC) farms, which generally have many computing nodes with a high number of processors. Large collaborations are working to use these resources in the most efficient way, compatibly with the constraints imposed by computing models (data distributed on the Grid, authentication, software dependencies, etc.). The aim of this thesis project is to develop a software framework that allows users to process a typical data analysis workflow of the ATLAS experiment on HPC systems. The developed analysis framework shall be deployed on the computing resources of the Open Physics Hub project and on the CINECA Marconi100 cluster, in view of the switch-on of the Leonardo supercomputer, foreseen in 2023.

analisi delle prestazioni del cluster risc-v: monte cimone

Gli sforzi di ricerca relativi all'High Performance Computing, nel corso degli anni, hanno prodotto risultati importanti inerenti all'incremento delle prestazioni sia in termini di numero di operazioni effettuate per periodo temporale, sia introducendo o migliorando algoritmi paralleli presenti in letteratura. Tali traguardi hanno comportato cambiamenti alla struttura interna delle macchine; si è assistito infatti ad un'evoluzione delle architetture dei processori utilizzati e all'impiego di GPU come risorse di calcolo aggiuntive. La conseguenza di un continuo incremento di prestazioni è quella di dover far fronte ad un grosso dispendio energetico, in quanto le macchine impiegate nell'HPC sono ideate per effettuare un'intensa attività di calcolo in un periodo di tempo molto prolungato; l'energia necessaria per alimentare ciascun nodo e dissipare il calore generato comporta costi elevati. Tra le varie soluzioni proposte per limitare il consumo di energia, quella che ha riscosso maggior interesse, sia a livello di studio che di mercato, è stata l'integrazione di CPU di tipologia RISC (Reduced Instruction Set Computer), in quanto capaci di ottenere prestazioni soddisfacenti con un impiego energetico inferiore rispetto alle CPU CISC (Complex Instruction Set Computer). In questa tesi è presentata l'analisi delle prestazioni di Monte Cimone, un cluster composto da 8 nodi di calcolo basati su architettura RISC-V e distribuiti in 4 piattaforme (\emph{blade}) dual-board. Verranno eseguiti dei benchmark che ci permetteranno di valutare: le prestazioni dello scambio di dati a lunga e corta distanza; le prestazioni nella risoluzione di problemi che presentano un principio di località spaziale ridotto; le prestazioni nella risoluzione di problemi su grafi e, nello specifico, ricerca in ampiezza e cammini minimi da sorgente singola.

SynBA: A contextualized Synonim-Based adversarial Attack for text classification

With the advent of high-performance computing devices, deep neural networks have gained a lot of popularity in solving many Natural Language Processing tasks. However, they are also vulnerable to adversarial attacks, which are able to modify the input text in order to mislead the target model. Adversarial attacks are a serious threat to the security of deep neural networks, and they can be used to craft adversarial examples that steer the model towards a wrong decision. In this dissertation, we propose SynBA, a novel contextualized synonym-based adversarial attack for text classification. SynBA is based on the idea of replacing words in the input text with their synonyms, which are selected according to the context of the sentence. We show that SynBA successfully generates adversarial examples that are able to fool the target model with a high success rate. We demonstrate three advantages of this proposed approach: (1) effective - it outperforms state-of-the-art attacks by semantic similarity and perturbation rate, (2) utility-preserving - it preserves semantic content, grammaticality, and correct types classified by humans, and (3) efficient - it performs attacks faster than other methods.

Chapel: analisi delle prestazioni su architetture a memoria distribuita

Nel presente elaborato si analizzeranno le prestazioni del linguaggio di programmazione parallela Chapel sul kernel Integer Sort di NAS Parallel Benchmarks. Questo algoritmo, a livello pratico, è utilizzato per studi o applicazioni sui metodi particellari. Saranno introdotti i concetti fondamentali di programmazione parallela e successivamente illustrate le principali caratteristiche di MPI e Chapel. Verranno poi approfonditi Integer Sort e i rispettivi dettagli implementativi, concludendo con un'analisi di prestazioni dei due linguaggi sul kernel preso in esame.

Energy consumption of parallel algorithms for solving linear systems on HPC architecture

Modern High-Performance Computing HPC systems are gradually increasing in size and complexity due to the correspondent demand of larger simulations requiring more complicated tasks and higher accuracy. However, as side effects of the Dennard’s scaling approaching its ultimate power limit, the efficiency of software plays also an important role in increasing the overall performance of a computation. Tools to measure application performance in these increasingly complex environments provide insights into the intricate ways in which software and hardware interact. The monitoring of the power consumption in order to save energy is possible through processors interfaces like Intel Running Average Power Limit RAPL. Given the low level of these interfaces, they are often paired with an application-level tool like Performance Application Programming Interface PAPI. Since several problems in many heterogeneous fields can be represented as a complex linear system, an optimized and scalable linear system solver algorithm can decrease significantly the time spent to compute its resolution. One of the most widely used algorithms deployed for the resolution of large simulation is the Gaussian Elimination, which has its most popular implementation for HPC systems in the Scalable Linear Algebra PACKage ScaLAPACK library. However, another relevant algorithm, which is increasing in popularity in the academic field, is the Inhibition Method. This thesis compares the energy consumption of the Inhibition Method and Gaussian Elimination from ScaLAPACK to profile their execution during the resolution of linear systems above the HPC architecture offered by CINECA. Moreover, it also collates the energy and power values for different ranks, nodes, and sockets configurations. The monitoring tools employed to track the energy consumption of these algorithms are PAPI and RAPL, that will be integrated with the parallel execution of the algorithms managed with the Message Passing Interface MPI.

Building a high-quality workforce : the Federal Career Intern Program /

"A report to the President and the Congress of the United States by the U.S. Merit Systems Protection Board."