Search

Hyperspectral image registration is a relevant task for real-time applications such as environmental disaster management or search and rescue scenarios. The HYFMGPU algorithm was proposed as a single-GPU high-performance solution, but the need for a distributed version has arisen due to the continuous evolution of sensors that generate images with finer spatial and spectral resolutions. In a previous work, we simplified the programming of the multi-device parts of an initial MPI+CUDA multi-GPU implementation of HYFMGPU by means of Hitmap, a library to ease the programming of parallel applications based on distributed arrays. The performance of that Hitmap version was assessed in a homogeneous GPU cluster. In this paper, we extend this implementation by means of new functionalities added to the latest version of Hitmap in order to support arbitrary load distributions for multi-node heterogeneous GPU clusters. Three different load balancing layouts are tested, which prove that selecting a proper layout affects the performance of the code and how this performance is correlated with the use of the GPUs available in the cluster.

In distributed-memory systems, data redistributions are operations that change the ownership and location of a selected subset of a data structure at runtime. They allow the improvement of the performance of parallel algorithms which operate on changing or partial domains, aiming to create a balanced workload among the active processes. To manually redistribute data is a cumbersome and error-prone task. In this paper, we present a method based on four combinable operators to redistribute partial domains selected by the programmer at runtime in an efficient and simple way. They abstract to the programmer the data-redistribution implementation details, such as the new mapping, relocation, and communication of the selected data. We also present the application of the proposed operators to a RayTracing application and to a significant part of STL (C++ Standard Template Library). Our experimental results show that our approach automatically generates a good load balance, which leads to performance improvements for generic data-distribution policies. It does not introduce significant performance overheads compared with tailored data redistributions directly programmed using MPI (Message Passing Interface), while it greatly reduces the code development effort.

Producción Científica, To predict the effectiveness of building evacuations is a very difficult task in the general case. In a previous work, the historical results of 47 evacuation drills in 15 different university buildings, both academic and residential, involving more than 19 000 persons, was analyzed, and a method based on dimensional analysis and statistical regression was proposed to give an estimation of the exit time in case of evacuation. Comparing this estimated exit time with the real values obtained in evacuation drills, more informed decisions on whether to invest in more training and/or preventive culture of the occupants or to invest in structural improvements of the buildings can be taken. In this work, we both propose a refinement of the method to calculate expected exit times, that leads to an even better adjustment between predictions and real-world results, and we use this refined model to predict the results of evacuations of a new building, whose use and characteristics are different from those previously studied, and whose data was provided by other authors in the bibliography. We show that there exists a correlation between the published results and the predictions generated by our model, both from a quantitative and qualitative point of view., Ministerio de Economía, Industria y Competitividad - Fondo Europeo de Desarrollo Regional (project TIN2017-88614-R), Junta de Castilla y León - Fondo Europeo de Desarrollo Regional (grants VA082P17 and VA226P20 and project INVESTUN-18-VA-0001)

Scientific applications are some of the most computationally demanding software pieces. Their core is usually a set of linear algebra operations, which may represent a significant part of the overall run-time of the application. BLAS libraries aim to solve this problem by exposing a set of highly optimized, reusable routines. There are several implementations specifically tuned for different types of computing platforms, including coprocessors. Some examples include the one bundled with the Intel MKL library, which targets Intel CPUs or Xeon Phi coprocessors, or the cuBLAS library, which is specifically designed for NVIDIA GPUs. Nowadays, computing nodes in many supercomputing clusters include one or more different coprocessor types. To fully exploit these platforms might require programs that can adapt at run-time to the chosen device type, hardwiring in the program the code needed to use a different library for each device type that can be selected. This also forces the programmer to deal with different interface particularities and mechanisms to manage the memory transfers of the data structures used as parameters. This paper presents a unified, performance-oriented, and portable interface for BLAS. This interface has been integrated into a heterogeneous programming model (Controllers) which supports groups of CPU cores, Xeon Phi accelerators, or NVIDIA GPUs in a transparent way. The contribution of this paper includes: An abstraction layer to hide programming differences between diverse BLAS libraries; new types of kernel classes to support the context manipulation of different external BLAS libraries; a new kernel selection policy that considers both programmer kernels and different external libraries; a complete new Controller library interface for the whole collection of BLAS routines. This proposal enables the creation of BLAS-based portable codes that can execute on top of different types of accelerators by changing a single initialization parameter. Our software internally exploits different preexisting and widely known BLAS library implementations, such as cuBLAS, MAGMA, or the one found in Intel MKL. It transparently uses the most appropriate library for the selected device. Our experimental results show that our abstraction does not introduce significant performance penalties, while achieving the desired portability.

The time needed to evacuate a building depends on many factors. Some are related to people's behavior, while others are related to the physical characteristics of the building. This paper analyzes the historical data of 47 evacuation drills in 15 different university buildings, both academic and residential, involving more than 19 000 persons. We propose the study of the data presented using a dimensionless analysis and statistical regression in order to give a prediction of the ratio between exit time and the number of people evacuated. The results obtained show that this approach could be a useful tool for comparing buildings of this type, and that it represents a promising research topic which can also be extended to other types of buildings.

Current HPC clusters are composed by several machines with different computation capabilities and different kinds and families of accelerators. Programming efficiently for these heterogeneous systems has become an important challenge. There are many proposals to simplify the programming and management of accelerator devices, and the hybrid programming, mixing accelerators and CPU cores. However, in many cases, portability compromises the efficiency on different devices, and there are details concerning the coordination of different types of devices that should still be tackled by the programmer. In this work, we introduce the Multi-Controller, an abstract entity implemented in a library that coordinates the management of heterogeneous devices, including accelerators with different capabilities and sets of CPU-cores. Our proposal improves state-of-the-art solutions, simplifying data partition, mapping and the transparent deployment of both, simple generic kernels portable across different device types, and specialized implementations defined and optimized using specific native or vendor programming models (such as CUDA for NVIDIA’s GPUs, or OpenMP for CPU-cores). The run-time system automatically selects and deploys the most appropriate implementation of each kernel for each device, managing data movements and hiding the launch details. The results of an experimental study with five study cases indicates that our abstraction allows the development of flexible and highly efficient programs that adapt to the heterogeneous environment.

Producción Científica, Current High Performance Computing (HPC) systems are typically built as interconnected clusters of shared-memory multicore computers. Several techniques to automatically generate parallel programs from high-level parallel languages or sequential codes have been proposed. To properly exploit the scalability of HPC clusters, these techniques should take into account the combination of data communication across distributed memory, and the exploitation of shared-memory models. In this paper, we present a new communication calculation technique to be applied across different SPMD (Single Program Multiple Data) code blocks, containing several uniform data access expressions. We have implemented this technique in Trasgo, a programming model and compilation framework that transforms parallel programs from a high-level parallel specification that deals with parallelism in a unified, abstract, and portable way. The proposed technique computes at runtime exact coarse-grained communications for distributed message-passing processes. Applying this technique at runtime has the advantage of being independent of compile-time decisions, such as the tile size chosen for each process. Our approach allows the automatic generation of pre-compiled multi-level parallel routines, libraries, or programs that can adapt their communication, synchronization, and optimization structures to the target system, even when computing nodes have different capabilities. Our experimental results show that, despite our runtime calculation, our approach can automatically produce efficient programs compared with MPI reference codes, and with codes generated with auto-parallelizing compilers., 2018-12-01, MICINN (Spain) and ERDF program of the European Union: HomProg-HetSys project (TIN2014-58876-P), CAPAP-H6 (TIN2016-81840- REDT), COST Program Action IC1305: Network for Sustainable Ultrascale Computing (NESUS), and by the computing facilities of Extremadura Research Centre for Advanced Technologies (CETA-CIEMAT), funded by the European Regional Development Fund (ERDF). CETACIEMAT belongs to CIEMAT and the Government of Spain.

Hyperspectral image registration is a relevant task for real-time applications like environmental disasters management or search and rescue scenarios. Traditional algorithms for this problem were not really devoted to real-time performance. The HYFMGPU algorithm arose as a high-performance GPU-based solution to solve such a lack. Nevertheless, a single-GPU solution is not enough, as sensors are evolving and then generating images with finer resolutions and wider wavelength ranges. An MPI+CUDA multi-GPU implementation of HYFMGPU was previously presented. However, this solution shows the programming complexity of combining MPI with an accelerator programming model. In this paper we present a new and more abstract programming approach for this type of applications, which provides a high efficiency while simplifying the programming of the multi-device parts of the code. The solution uses Hitmap, a library to ease the programming of parallel applications based on distributed arrays. It uses a more algorithm-oriented approach than MPI, including abstractions for the automatic partition and mapping of arrays at runtime with arbitrary granularity, as well as techniques to build flexible communication patterns that transparently adapt to the data partitions. We show how these abstractions apply to this application class. We present a comparison of development effort metrics between the original MPI implementation and the one based on Hitmap, with reductions of up to 95% for the Halstead score in specific work redistribution steps. We finally present experimental results showing that these abstractions are internally implemented in a high efficient way that can reduce the overall performance time in up to 37% comparing with the original MPI implementation.

Thread-Level Speculation (TLS) is a promising technique that allows the parallel execution of sequential code without relying on a prior, compile-time-dependence analysis. In this work, we introduce the technique, present a taxonomy of TLS solutions, and summarize and put into perspective the most relevant advances in this field.

Parallelization of sequential applications requires extracting information about the loops and how their variables are accessed, and afterwards, augmenting the source code with extra code depending on such information. In this paper we propose a framework that avoids such an error-prone, time-consuming task. Our solution leverages the compile-time information extracted from the source code to classify all variables used inside each loop according to their accesses. Then, our system, called BFCA+, automatically instruments the source code with the necessary OpenMP directives and clauses to allow its parallel execution, using the standard shared and private clauses for variable classification. The framework is also capable of instrumenting loops for speculative parallelization, with the help of the ATLaS runtime system, that defines a new speculative clause to point out those variables that may lead to a dependency violation. As a result, the target loop is guaranteed to correctly run in parallel, ensuring that its execution follows sequential semantics even in the presence of dependency violations. Our experimental evaluation shows that the framework not only saves development time, but also leads to a faster code than the one manually parallelized.

OpenMP directives are the de-facto standard for shared-memory parallel programming. However, OpenMP does not guarantee the correctness of the parallel execution of a given loop if runtime data dependences arise. Consequently, many highly-parallel regions cannot be safely parallelized with OpenMP due to the possibility of a dependence violation. In this paper, we propose to augment OpenMP capabilities, by adding thread-level speculation (TLS) support. Our contribution is threefold. First, we have defined a new speculative clause for variables inside parallel loops. This clause ensures that all accesses to these variables will be carried out according to sequential semantics. Second, we have created a new, software-based TLS runtime library to ensure correctness in the parallel execution of OpenMP loops that include speculative variables. Third, we have developed a new GCC plugin, which seamlessly translates our OpenMP speculative clause into calls to our TLS runtime engine. The result is the ATLaS C Compiler framework, which takes advantage of TLS techniques to expand OpenMP functionalities, and guarantees the sequential semantics of any parallelized loop.

Producción Científica, Dataflow programming consists in developing a program by describing its sequential stages and the interactions between them. The runtime systems supporting this kind of programming are responsible for exploiting the parallelism by concurrently executing the different stages as soon as their dependencies are met. In this paper we introduce a new parallel programming model and framework based on the dataflow paradigm. It presents a new combination of features that allows to easily map programs to shared or distributed memory, exploiting data locality and affinity to obtain the same performance than optimized coarse-grain MPI programs. These features include: It is a unique one-tier model that supports hybrid shared- and distributed-memory systems with the same abstractions; it can express activities arbitrarily linked, including non-nested cycles; it uses internally a distributed work-stealing mechanism to allow Multiple-Producer/Multiple-Consumer configurations; and it has a runtime mechanism for the reconfiguration of the dependences and communication channels which also allows the creation of task-to-task data affinities. We present an evaluation using examples of different classes of applications. Experimental results show that programs generated using this framework deliver good performance in hybrid distributed- and shared-memory environments, with a similar development effort as other dataflow programming models oriented to shared-memory., 2019-01-01, MICINN (Spain) and ERDF program of the European Union: HomProg-HetSys project (TIN2014-58876- P), PCAS project (TIN2017-88614-R), CAPAP-H6 (TIN2016-81840-REDT), and COST Program Action IC1305: Network for Sustainable Ultrascale Com- puting (NESUS). By Junta de Castilla y Le on, project PROPHET (VA082P17). And by the computing facilities of Extremadura Research Centre for Advanced Technologies (CETA- CIEMAT), funded by the European Regional Develop- ment Fund (ERDF). CETA-CIEMAT belongs to CIEMAT and the Govern- ment of Spain.

Pursuing a college degree is a task that requires a great amount of time and effort. Universities are facing a big challenge to attract students and keep them motivated. The gamification of education is a practice that expects to increase the students' engagement, which in turn increases learning outcomes. Nevertheless, obtaining beneficial results from gamification requires educators to mold the teaching to include this new practice, usually involving a lot of effort. In this chapter, the authors present a new software tool developed to encourage gamification dynamics, and they describe their experience using this tool in a Parallel Programming course. The chapter describes the structure of the course, the different proposed activities, the organization of hardware resources, the design of the developed software tool, and an evaluation of the gamified course. The results show that the use of gamification techniques has been a great success. The authors have had a very positive response from their students, and there has been also a big percentage of passing students.

Waste disposal and recycling is becoming one of the main problems in Western countries. Improving both recycling culture among citizenship and waste collection and treatment logistics is critical to augment the percentage of waste being recycled. In this paper we present STERLING, an initiative that aims to help in both fields. STERLING is a framework composed by a low-cost, low-energy sensor installed in recycling containers to measure fi ll level and other physical parameters. The sensor is activated magnetically each time the container lid is opened by a user. Instead of directly sending this information to a cloud-based server, our sensor broadcasts a Bluetooth Low Energy (BLE) packet to the surrounding area. An App running in the mobile phone of the user performs two actions: To capture this information, re-sending it to the cloud-based server, and to assign credit to this particular user for having used the recycling container. In this way, users are rewarded for using the container, and the infrastructure bene ts from cost-free communications from the container to the server. In this paper we will describe our idea in detail, showing how it can be used to develop a rewarding schema that encourages recycling., Universidad de Valladolid (Consejería de Educación of Junta de Castilla y León, Ministerio de Economía, Industria y Competitividad of Spain, and European Regional Development Fund (ERDF) program: Project PCAS (TIN2017-88614-R), Project PROPHET (VA082P17) and CAPAP-H6 network (TIN2016-81840-REDT).

During the first decade of the twenty-first century, the advent of multicore processing reached its maturity level, with the help of shared-memory programming models such as OpenMP [1], that allows to parallelize both legacy and new C and Fortran applications in a shared-memory environments. Meanwhile, message-passing programming models such as MPI [2] allowed to aggregate multicore systems in larger clusters, which dominated the TOP 500 supercomputing list [3]. However, at that time parallel computing seemed to face some limits that were hard to overcome. Physical limits prevented clock frequencies to increase, and the Law of Diminishing Returns reduced the usefulness of keep adding cores to a multiprocessor. Suddenly, the advent of GPU computing changed the game once again. Being initially developed as a way to accelerate graphical processing, GPUs were reused to speed up certain types of calculations that needed a similar processing to an entire set of independent elements. CUDA [4] programming model allowed programmers to translate to the GPU world many applications, and the TOP 500 list started to show more and more heterogeneous systems that incorporated accelerators to their cluster nodes., Este trabajo forma parte del proyecto de investigación PCAS Grant TIN2017-88614-R y la Junta de Castilla y León, proyecto PROPHET, VA082P17.