Interactive computing notebooks, such as Jupyter notebooks, have become a popular tool for developing and improving data-driven models. Such notebooks tend to be executed either in the user’s own machine or in a cloud environment, having drawbacks and benefits in both approaches. This paper presents a solution developed as a Jupyter extension that automatically selects which cells, as well as in which scenarios, such cells should be migrated to a more suitable platform for execution. We describe how we reduce the execution state of the notebook to decrease migration time and we explore the knowledge of user interactivity patterns with the notebook to determine which blocks of cells should be migrated. Using notebooks from Earth science (remote sensing), image recognition, and hand written digit identification (machine learning), our experiments show notebook state reductions of up to 55× and migration decisions leading to performance gains of up to 3.25× when the user interactivity with the notebook is taken into consideration.

Several industries rely on accurate and efficient processing of 3D spatial queries over increasingly large datasets for decision optimization and exploration purposes. Examples include clinical diagnosis supported by 3D imaging of human tissues, numerical simulation of aerodynamics during the design of aircraft and vehicles, and the search for profitable deposits of minerals, oil and gas guided by 3D maps extrapolated from dense collections of rock samples. Despite the clear demand for spatial data processing in 3D space, existing systems supporting these organizations are scarce and scale poorly with data volume. This paper presents a GPU-based acceleration engine for SQL database management systems that can serve spatial queries orders of magnitude faster than solutions based on traditional software stacks.

Package management is instrumental for programming languages and operating systems, and yet it is neglected by both areas as an implementation detail. For this reason, it lacks the same kind of conceptual organization: we lack terminology to classify them or to reason about their design trade-offs. In this paper, we share our experience in both OS and language-specific package manager development, categorizing families of package managers and discussing their design implications beyond particular implementations. We also identify possibilities in the still largely unexplored area of package manager interoperability.

Online subtitle databases allow users to easily find subtitle documents in multiple languages for thousands of films and TV series episodes. However, getting the subtitle document that gives satisfactory synchronization on the first attempt is like hitting the jackpot. The truth is that this process often involves a lot of trial- and-error because multiple versions of subtitle documents have distinct synchronization references, given that they are targeted at variations of the same audiovisual content. Building on our previous efforts to address this problem, in this paper we formalize and validate a two-phase subtitle synchronization framework. The benefit over current approaches lays in the usage of audio fingerprint annotations generated from the base audio signal as second-level synchronization anchors. This way, we allow the media player to dynamically fix during playback the most common cases of subtitle synchronization misalignment that compromise users' watching experience. Results from our evaluation process indicate that our framework has minimal impact on existing subtitle documents and formats as well as on the playback performance.

Numerous problems in science and engineering involve discretizing the problem domain as a regular structured grid and make use of domain decomposition techniques to obtain solutions faster using high performance computing. However, the load imbalance of the workloads among the various processing nodes can cause severe degradation in application performance. This problem is exacerbated for the case when the computational workload is non-uniform and the processing nodes have varying computational capabilities. In this paper, we present novel local search algorithms for regular partitioning of a structured mesh to heterogeneous compute nodes in a distributed setting. The algorithms seek to assign larger workloads to processing nodes having higher computation capabilities while maintaining the regular structure of the mesh in order to achieve a better load balance. We also propose a distributed memory (MPI) parallelization architecture that can be used to achieve a parallel implementation of scientific modelling software requiring structured grids on heterogeneous processing resources involving CPUs and GPUs. Our implementation can make use of the available CPU cores and multiple GPUs of the underlying platform simultaneously. Empirical evaluation on real world flood modelling domains on a heterogeneous architecture comprising of multicore CPUs and GPUs suggests that the proposed partitioning approach can provide a performance improvement of up to 8× over a naive uniform partitioning.

Anyone who ever downloaded subtitle files from the Internet has faced problems synchronizing them with the associated media files. Even with the efforts of communities on reviewing user-contributed subtitles and with mechanisms in movie players to automate the discovery of subtitles for a given media, users still face lip synchronization issues. In this work we conduct a study on several subtitle files associated with popular movies and TV series and analyze their differences. Based on that, we propose a two-phase subtitle synchronization method that annotates subtitles with audio fingerprints, which serve as synchronization anchors to the media player. Preliminary results obtained with our prototype suggest that our technique is effective and has minimal impact on the extension of subtitle formats and on media playback performance.

For a long time, tapes have had a single logical partition that was efficiently operated by dedicated software in batch mode. Nowadays, after the introduction of the LTO 5 standard, tapes support logical partitioning and, with the arrival of the Linear Tape File System (LTFS), their contents are exposed through the file system interface as regular files and directories. As a consequence, tape medium can now be accessed by concurrent processes that may create files in parallel. This creates two potential problems: interleaved data blocks and expensive partition switches. This paper presents the design and implementation of Unified, an I/O scheduler for LTFS that addresses these problems. We observe the effectiveness of delayed writes on decreased file fragmentation and the reduction of partition switches with the use of buffering and of redundant file copies. We also find out that software-based read prefetching, commonly used to manage disk devices, does not improve read times on tapes, but rather introduces potential overhead. With the techniques described in this paper, the Unified scheduler allows tape operations to be performed close to the raw hardware speed.

Accurate and timely flood forecasts are becoming highly essential due to the increased incidence of flood related disasters over the last few years. Such forecasts require a high resolution integrated flood modeling approach. In this paper, we present an integrated flood forecasting system with an automated workflow over the weather modeling, surface runoff estimation and water routing components. We primarily focus on the water routing process which is the most compute intensive phase and present two parallelization strategies to scale it up to large grid sizes. Specifically, we employ nature-inspired decomposition of a simulation domain into watershed basins and propose a master slave model of parallelization for distributed processing of the basins. We also propose an intra-basin shared memory parallelization approach using OpenMP. Empirical evaluation of the proposed parallelization strategies indicates a potential for high speedups for certain types of scenarios (e.g., speedup of 13× with 16 threads using OpenMP parallelization for the large Rio de Janeiro basin).

While digital video cameras have existed for over two decades digital video cassettes are still the primary storage medium in professional video archives. One of the major inhibitors in the transition to file-based workflows and media archives is the lack of an affordable, portable and archive compatible storage medium for the vast amounts of content produced.

While there are many financial and practical reasons to prefer tape storage over disk for various applications, the difficultly of using tape in a general way is a major inhibitor to its wider usage. We present a file system that takes advantage of a new generation of tape hardware to provide efficient access to tape using standard, familiar system tools and interfaces. The Linear Tape File System (LTFS) makes using tape as easy, flexible, portable, and intuitive as using other removable and sharable media, such as a USB drive.