Search

Flash-based SSDs have become well established in the storage market, replacing magnetic disks in both enterprise and consumer computer systems. The performance characteristics of these new devices have prompted a considerable amount of research that aims at developing efficient data access methods. Early works attempted to reduce the expensive random writes, exploiting logging and batch write techniques, while more recent ones addressed query processing, taking advantage of the high internal parallelism of SSDs. 3D XPoint is a new nonvolatile memory technology that has emerged recently, featuring smaller access times and higher durability compared with flash. It is available both as block addressable secondary storage and as byte addressable persistent main memory. However, the high cost of 3D XPoint prevents, for the moment, its adoption in large scales. This renders hybrid storage systems utilizing NAND flash and 3D XPoint a sufficient alternative. In this work, we propose HyR-tree, a hybrid variant of R-tree that persists a part of the tree in the high performing 3D XPoint storage. HyR-tree identifies repeated access pattern to the data and uses these patterns to locate the most important nodes. The importance of a node is determined by the performance gain that derives from its placement within a 3D XPoint-based device. We experimentally evaluated HyR-tree using real devices and four different datasets. The acquired results show that our proposal achieves significant performance gains up to 40% in tree construction and up to 56% in range queries. [ABSTRACT FROM AUTHOR]

We present the full public release of all data from the TNG100 and TNG300 simulations of the IllustrisTNG project. IllustrisTNG is a suite of large volume, cosmological, gravo-magnetohydrodynamical simulations run with the moving-mesh code Arepo. TNG includes a comprehensive model for galaxy formation physics, and each TNG simulation self-consistently solves for the coupled evolution of dark matter, cosmic gas, luminous stars, and supermassive black holes from early time to the present day, z = 0 . Each of the flagship runs—TNG50, TNG100, and TNG300—are accompanied by halo/subhalo catalogs, merger trees, lower-resolution and dark-matter only counterparts, all available with 100 snapshots. We discuss scientific and numerical cautions and caveats relevant when using TNG. The data volume now directly accessible online is ∼750 TB, including 1200 full volume snapshots and ∼80,000 high time-resolution subbox snapshots. This will increase to ∼1.1 PB with the future release of TNG50. Data access and analysis examples are available in IDL, Python, and Matlab. We describe improvements and new functionality in the web-based API, including on-demand visualization and analysis of galaxies and halos, exploratory plotting of scaling relations and other relationships between galactic and halo properties, and a new JupyterLab interface. This provides an online, browser-based, near-native data analysis platform enabling user computation with local access to TNG data, alleviating the need to download large datasets. [ABSTRACT FROM AUTHOR]