Subscribe to the gold package and get unlimited access to Shamra Academy
Register a new userRealizing Fast, Scalable and Reliable Scientific Computations in Grid Environments
72
0
0.0
(
0
)
Added by
Ioan Raicu
Publication date
2008
fields
Informatics Engineering
and research's language is
English
Ask ChatGPT about the research
No Arabic abstract
The practical realization of managing and executing large scale scientific computations efficiently and reliably is quite challenging. Scientific computations often involve thousands or even millions of tasks operating on large quantities of data, such data are often diversely structured and stored in heterogeneous physical formats, and scientists must specify and run such computations over extended periods on collections of compute, storage and network resources that are heterogeneous, distributed and may change constantly. We present the integration of several advanced systems: Swift, Karajan, and Falkon, to address the challenges in running various large scale scientific applications in Grid environments. Swift is a parallel programming tool for rapid and reliable specification, execution, and management of large-scale science and engineering workflows. Swift consists of a simple scripting language called SwiftScript and a powerful runtime system that is based on the CoG Karajan workflow engine and integrates the Falkon light-weight task execution service that uses multi-level scheduling and a streamlined dispatcher. We showcase the scalability, performance and reliability of the integrated system using application examples drawn from astronomy, cognitive neuroscience and molecular dynamics, which all comprise large number of fine-grained jobs. We show that Swift is able to represent dynamic workflows whose structures can only be determined during runtime and reduce largely the code size of various workflow representations using SwiftScript; schedule the execution of hundreds of thousands of parallel computations via the Karajan engine; and achieve up to 90% reduction in execution time when compared to traditional batch schedulers.
rate research
Read More
Rapid growth in scientific data and a widening gap between computational speed and I/O bandwidth makes it increasingly infeasible to store and share all data produced by scientific simulations. Instead, we need methods for reducing data volumes: ideally, methods that can scale data volumes adaptively so as to enable negotiation of performance and fidelity tradeoffs in different situations. Multigrid-based hierarchical data representations hold promise as a solution to this problem, allowing for flexible conversion between different fidelities so that, for example, data can be created at high fidelity and then transferred or stored at lower fidelity via logically simple and mathematically sound operations. However, the effective use of such representations has been hindered until now by the relatively high costs of creating, accessing, reducing, and otherwise operating on such representations. We describe here highly optimized data refactoring kernels for GPU accelerators that enable efficient creation and manipulation of data in multigrid-based hierarchical forms. We demonstrate that our optimized design can achieve up to 264 TB/s aggregated data refactoring throughput -- 92% of theoretical peak -- on 1024 nodes of the Summit supercomputer. We showcase our optimized design by applying it to a large-scale scientific visualization workflow and the MGARD lossy compression software.
Scientific computing sometimes involves computation on sensitive data. Depending on the data and the execution environment, the HPC (high-performance computing) user or data provider may require confidentiality and/or integrity guarantees. To study the applicability of hardware-based trusted execution environments (TEEs) to enable secure scientific computing, we deeply analyze the performance impact of AMD SEV and Intel SGX for diverse HPC benchmarks including traditional scientific computing, machine learning, graph analytics, and emerging scientific computing workloads. We observe three main findings: 1) SEV requires careful memory placement on large scale NUMA machines (1$times$$-$3.4$times$ slowdown without and 1$times$$-$1.15$times$ slowdown with NUMA aware placement), 2) virtualization$-$a prerequisite for SEV$-$results in performance degradation for workloads with irregular memory accesses and large working sets (1$times$$-$4$times$ slowdown compared to native execution for graph applications) and 3) SGX is inappropriate for HPC given its limited secure memory size and inflexible programming model (1.2$times$$-$126$times$ slowdown over unsecure execution). Finally, we discuss forthcoming new TEE designs and their potential impact on scientific computing.
In the current era of Big Data, data engineering has transformed into an essential field of study across many branches of science. Advancements in Artificial Intelligence (AI) have broadened the scope of data engineering and opened up new applications in both enterprise and research communities. Aggregations (also termed reduce in functional programming) are an integral functionality in these applications. They are traditionally aimed at generating meaningful information on large data-sets, and today, they are being used for engineering more effective features for complex AI models. Aggregations are usually carried out on top of data abstractions such as tables/ arrays and are combined with other operations such as grouping of values. There are frameworks that excel in the said domains individually. But, we believe that there is an essential requirement for a data analytics tool that can universally integrate with existing frameworks, and thereby increase the productivity and efficiency of the entire data analytics pipeline. Cylon endeavors to fulfill this void. In this paper, we present Cylons fast and scalable aggregation operations implemented on top of a distributed in-memory table structure that universally integrates with existing frameworks.
Designing efficient and scalable sparse linear algebra kernels on modern multi-GPU based HPC systems is a daunting task due to significant irregular memory references and workload imbalance across the GPUs. This is particularly the case for Sparse Triangular Solver (SpTRSV) which introduces additional two-dimensional computation dependencies among subsequent computation steps. Dependency information is exchanged and shared among GPUs, thus warrant for efficient memory allocation, data partitioning, and workload distribution as well as fine-grained communication and synchronization support. In this work, we demonstrate that directly adopting unified memory can adversely affect the performance of SpTRSV on multi-GPU architectures, despite linking via fast interconnect like NVLinks and NVSwitches. Alternatively, we employ the latest NVSHMEM technology based on Partitioned Global Address Space programming model to enable efficient fine-grained communication and drastic synchronization overhead reduction. Furthermore, to handle workload imbalance, we propose a malleable task-pool execution model which can further enhance the utilization of GPUs. By applying these techniques, our experiments on the NVIDIA multi-GPU supernode V100-DGX-1 and DGX-2 systems demonstrate that our design can achieve on average 3.53x (up to 9.86x) speedup on a DGX-1 system and 3.66x (up to 9.64x) speedup on a DGX-2 system with 4-GPUs over the Unified-Memory design. The comprehensive sensitivity and scalability studies also show that the proposed zero-copy SpTRSV is able to fully utilize the computing and communication resources of the multi-GPU system.
Distributed approaches based on the map-reduce programming paradigm have started to be proposed in the bioinformatics domain, due to the large amount of data produced by the next-generation sequencing techniques. However, the use of map-reduce and related Big Data technologies and frameworks (e.g., Apache Hadoop and Spark) does not necessarily produce satisfactory results, in terms of both efficiency and effectiveness. We discuss how the development of distributed and Big Data management technologies has affected the analysis of large datasets of biological sequences. Moreover, we show how the choice of different parameter configurations and the careful engineering of the software with respect to the specific framework under consideration may be crucial in order to achieve good performance, especially on very large amounts of data. We choose k-mers counting as a case study for our analysis, and Spark as the framework to implement FastKmer, a novel approach for the extraction of k-mer statistics from large collection of biological sequences, with arbitrary values of k. One of the most relevant contributions of FastKmer is the introduction of a module for balancing the statistics aggregation workload over the nodes of a computing cluster, in order to overcome data skew while allowing for a fully exploitation of the underly- ing distributed architecture. We also present the results of a comparative experimental analysis showing that our approach is currently the fastest among the ones based on Big Data technologies, while exhibiting a very good scalability. We provide evidence that the usage of technologies such as Hadoop or Spark for the analysis of big datasets of biological sequences is productive only if the architectural details and the peculiar aspects of the considered framework are carefully taken into account for the algorithm design and implementation.
Log in to be able to interact and post comments
comments
Fetching comments
Sign in to be able to follow your search criteria
