Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “Distributed Memory Parallel”

Search indexed NASA NTRS and DOE OSTI research on propulsion, heat transfer, battery materials and energy systems. Follow report and document links to the original sources.

Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.

At least 469 records · Page 26

On Data Transfers Over Wide-Area Dedicated Connections

Dedicated wide-area network connections are employed in big data and high-performance computing scenarios, since the absence of cross-traffic promises to make it easier to analyze and optimize data transfers over them. However, nonlinear transport dynamics and end-system complexity due to multi-core hosts and distributed file systems make these tasks surprisingly challenging. We present an overview of methods to analyze memory and disk file transfers using extensive measurements over 10 Gbps physical and emulated connections with 0–366 ms round trip times (RTTs). For memory transfers, we derive performance profiles of TCP and UDT throughput as a function of RTT, which show concave regions in contrast to entirely convex regions predicted by previous models. These highly desirable concave regions can be expanded by utilizing large buffers and more parallel flows. We also present Poincar´e maps and Lyapunov exponents of TCP and UDT throughputtraces that indicate complex throughput dynamics. For disk file transfers, we show that throughput can be optimized using a combination of parallel I/O and network threads under direct I/O mode. Our initial throughput measurements of Lustre filesystems mounted over long-haul connections using LNet routers show convex profiles indicative of I/O limits.

Rao, Nageswara S.↗

Fast analysis of molecular dynamics trajectories with graphics processing units-Radial distribution function histogramming

The calculation of radial distribution functions (RDFs) from molecular dynamics trajectory data is a common and computationally expensive analysis task. The rate limiting step in the calculation of the RDF is building a histogram of the distance between atom pairs in each trajectory frame. Here we present an implementation of this histogramming scheme for multiple graphics processing units (GPUs). The algorithm features a tiling scheme to maximize the reuse of data at the fastest levels of the GPU's memory hierarchy and dynamic load balancing to allow high performance on heterogeneous configurations of GPUs. Several versions of the RDF algorithm are presented, utilizing the specific hardware features found on different generations of GPUs. We take advantage of larger shared memory and atomic memory operations available on state-of-the-art GPUs to accelerate the code significantly. The use of atomic memory operations allows the fast, limited-capacity on-chip memory to be used much more efficiently, resulting in a fivefold increase in performance compared to the version of the algorithm without atomic operations. The ultimate version of the algorithm running in parallel on four NVIDIA GeForce GTX 480 (Fermi) GPUs was found to be 92 times faster than a multithreaded implementation running on an Intel Xeon 5550 CPU. On this multi-GPU hardware, the RDF between two selections of 1,000,000 atoms each can be calculated in 26.9 s per frame. The multi-GPU RDF algorithms described here are implemented in VMD, a widely used and freely available software package for molecular dynamics visualization and analysis.

97 MATHEMATICS AND COMPUTING↗

Integrating PGAS and MPI-based Graph Analysis

This project demonstrates that Chapel programs can interface with MPI-based libraries written in C++ without storing multiple copies of shared data. Chapel is a language for productive parallel computing using global address spaces (PGAS). We identified two approaches to interface Chapel code with the MPI-based Grafiki and Trilinos libraries. The first uses a single Chapel executable to call a C function that interacts with the C++ libraries. The second uses the mmap function to allow separate executables to read and write to the same block of memory on a node. We also encapsulated the second approach in Docker/Singularity containers to maximize ease of use. Comparisons of the two approaches using shared and distributed memory installations of Chapel show that both approaches provide similar scalability and performance.

97 MATHEMATICS AND COMPUTING↗

Mads.jl

Mads.jl (Model analysis and decision support in Julia) is a code that streamlines the process of using data and models for analysis and decision support. It is based on another open-source code developed at LANL and written in C/C++ (MADS; http://mads.lanl.gov; LA-CC-11- 035). Mads.jl can work with external models of arbitrary complexity as well as built-in models of flow and transport in porous media. It enables a number of data- and model-based analyses including model calibration, sensitivity analysis, uncertainty quantification, and decision analysis. The code also can use a series of alternative adaptive computational techniques for Bayesian sampling, Monte Carlo, and Bayesian Information-Gap Decision Theory. The code is implemented in the Julia programming language, and has high-performance (parallel) and memory management capabilities. The code uses a series of third party modules developed by others. The code development will also include contributions to the existing third party modules written in Julia; this contributions will be important for the efficient implementation of the algorithm used by Mads.jl. The code also uses a series of LANL developed modules that are developed by Dan O'Malley; these modules will be also a part of the Mads.jl release. Mads.jl will be released under GPL V3 license. The code will be distributed as a Git repo at gitlab.com and github.com. Mads.jl manual and documentation will be posted at madsjulia.lanl.gov.

Vesselinov, Velimir↗

Mads.jl

Mads.jl (Model analysis and decision support in Julia) is a code that streamlines the process of using data and models for analysis and decision support. It is based on another open-source code developed at LANL and written in C/C++ (MADS; http://mads.lanl.gov; LA-CC-11- 035). Mads.jl can work with external models of arbitrary complexity as well as built-in models of flow and transport in porous media. It enables a number of data- and model-based analyses including model calibration, sensitivity analysis, uncertainty quantification, and decision analysis. The code also can use a series of alternative adaptive computational techniques for Bayesian sampling, Monte Carlo, and Bayesian Information-Gap Decision Theory. The code is implemented in the Julia programming language, and has high-performance (parallel) and memory management capabilities. The code uses a series of third party modules developed by others. The code development will also include contributions to the existing third party modules written in Julia; this contributions will be important for the efficient implementation of the algorithm used by Mads.jl. The code also uses a series of LANL developed modules that are developed by Dan O'Malley; these modules will be also a part of the Mads.jl release. Mads.jl will be released under GPL V3 license. The code will be distributed as a Git repo at gitlab.com and github.com. Mads.jl manual and documentation will be posted at madsjulia.lanl.gov.

Vesselinov, Velimir↗

Forward-Weighted CADIS Method for Variance Reduction of Monte Carlo Reactor Analyses

Current state-of-the-art tools and methods used to perform 'real' commercial reactor analyses use high-fidelity transport codes to produce few-group parameters at the assembly level for use in low-order methods applied at the core level. Monte Carlo (MC) methods, which allow detailed and accurate modeling of the full geometry and energy details and are considered the 'gold standard' for radiation transport solutions, are playing an ever-increasing role in correcting and/or verifying the several-decade-old methodology used in current practice. However, the prohibitive computational requirements associated with obtaining fully converged system-wide solutions restrict the role of MC to benchmarking deterministic results at a limited number of state-points for a limited number of relevant quantities. A goal of current research at Oak Ridge National Laboratory (ORNL) is to change this paradigm by enabling the direct use of MC for full-core reactor analyses. The most significant of the many technical challenges that must be overcome is the slow non-uniform convergence of system-wide MC estimates and the memory requirements associated with detailed solutions throughout a reactor (problems involving hundreds of millions of different material and tally regions due to fuel irradiation, temperature distributions, and the needs associated with multi-physics code coupling). To address these challenges, research has focused on development in the following two areas: (1) a hybrid deterministic/MC method for determining high-precision fluxes throughout the problem space in k-eigenvalue problems and (2) an efficient MC domain-decomposition algorithm that partitions the problem phase space onto multiple processors for massively parallel systems, with statistical uncertainty estimation. The focus of this paper is limited to the first area mentioned above. It describes the FW-CADIS method applied to variance reduction of MC reactor analyses and provides initial results for calculating group-wise fluxes throughout a generic 2-D pressurized water reactor (PWR) quarter core model.

21 SPECIFIC NUCLEAR REACTORS AND ASSOCIATED PLANTS↗

Multi-fault Tolerance for Cartesian Data Distributions

Faults are expected to play an increasingly important role in how algorithms and applications are designed to run on future extreme-scale sys- tems. Algorithm-based fault tolerance (ABFT) is a promising approach that involves modications to the algorithm to recover from faults with lower over- heads than replicated storage and a signicant reduction in lost work compared to checkpoint-restart techniques. Fault-tolerant linear algebra (FTLA) algo- rithms employ additional processors that store parities along the dimensions of a matrix to tolerate multiple, simultaneous faults. Existing approaches as- sume regular data distributions (blocked or block-cyclic) with the failures of each data block being independent. To match the characteristics of failures on parallel computers, we extend these approaches to mapping parity blocks in several important ways. First, we handle parity computation for generalized Cartesian data distributions with each processor holding arbitrary subsets of blocks in a Cartesian-distributed array. Second, techniques to handle corre- lated failures, i.e., multiple processors that can be expected to fail together, are presented. Third, we handle the colocation of parity blocks with the data blocks and do not require them to be on additional processors. Several al- ternative approaches, based on graph matching, are presented that attempt to balance the memory overhead on processors while guaranteeing the same fault tolerance properties as existing approaches that assume independent fail- ures on regular blocked data distributions. The evaluation of these algorithms demonstrates that the additional desirable properties are provided by the pro- posed approach with minimal overhead.

Ali, Nawab↗

Final Report for Project DE-FC02-06ER25755 [Pmodels2]

In this report, we describe the research accomplished by the OSU team under the Pmodels2 project. The team has worked on various angles: designing high performance MPI implementations on modern networking technologies (Mellanox InfiniBand (including the new ConnectX2 architecture and Quad Data Rate), QLogic InfiniPath, the emerging 10GigE/iWARP and RDMA over Converged Enhanced Ethernet (RoCE) and Obsidian IB-WAN), studying MPI scalability issues for multi-thousand node clusters using XRC transport, scalable job start-up, dynamic process management support, efficient one-sided communication, protocol offloading and designing scalable collective communication libraries for emerging multi-core architectures. New designs conforming to the Argonne’s Nemesis interface have also been carried out. All of these above solutions have been integrated into the open-source MVAPICH/MVAPICH2 software. This software is currently being used by more than 2,100 organizations worldwide (in 71 countries). As of January ’14, more than 200,000 downloads have taken place from the OSU Web site. In addition, many InfiniBand vendors, server vendors, system integrators and Linux distributors have been incorporating MVAPICH/MVAPICH2 into their software stacks and distributing it. Several InfiniBand systems using MVAPICH/MVAPICH2 have obtained positions in the TOP500 ranking of supercomputers in the world. The latest November ’13 ranking include the following systems: 7th ranked Stampede system at TACC with 462,462 cores; 11th ranked Tsubame 2.5 system at Tokyo Institute of Technology with 74,358 cores; 16th ranked Pleiades system at NASA with 81,920 cores; Work on PGAS models has proceeded on multiple directions. The Scioto framework, which supports task-parallelism in one-sided and global-view parallel programming, has been extended to allow multi-processor tasks that are executed by processor groups. A quantum Monte Carlo application is being ported onto the extended Scioto framework. A public release of Global Trees (GT) has been made, along with the Global Chunks (GC) framework on which GT is built. The Global Chunks (GC) layer is also being used as the basis for the development of a higher level Global Graphs (GG) layer. The Global Graphs (GG) system will provide a global address space view of distributed graph data structures on distributed memory systems.

97 MATHEMATICS AND COMPUTING↗

CUDA Computation of the Feynman Distribution

In 2006, NVIDIA Corporation introduced the Computer Unified Device Architecture (CUDA) for C, C++, and Fortran programming languages. The CUDA parallel computing platform uses the Graphics Processing Unit (GPU) of NVIDIA graphics card, and it is designed for parallel computations. In these computations, the same algorithms are executed on the data in parallel. The GPU contains many Streaming Multiprocessors (SM) that can independently execute blocks of threads. A thread is a sequence of serial programmed instructions. The SM executes threads in groups of 32 and a group of 32 threads is called warp. The threads within a block can be synchronized and have access to a shared memory that is visible only to them. Conversely, blocks of threads cannot be synchronized and have access only to the global memory of the graphics card. This memory is called device memory and is independent of the host memory controlled by the Central Processing Unit (CPU). A single thread has access to a local memory (register) that is not visible to any other thread. The part of the program that is executed on the graphics card is called kernel. Each thread of the kernel has a unique identification index that is accessible by the in-built variable threadIdx; the latter is a 3-component vector. A threads block can contain up to 1024 threads. Blocks of threads have a unique identification index that is accessible by the in-built variable blockIdx; the latter is a 3-component vector. The in-built variable blockDim contains the number of threads per block; this variable is a 3-component vector. Threads and blocks represent different levels of parallelism; a kernel is executed as a grid of blocks of threads. Figure 1 illustrates a grid of thread blocks for a GPU with 2 SM. A SM can execute multiple blocks of threads in parallel; a block of threads is executed on a single SM. A GPU can execute multiple kernels concurrently. This work has used the Quadro M5000 graphics card with the following technical specifications: - 2048 cores with 861 MHz frequency, - 16 streaming multiprocessors, - 5.2 trillion transistors, - 8.2 Gb device memory, - 64 kb register memory (per thread), - 48 kb shared memory (per block) - 211.6 GB/s memory bandwidth, and - 3.5 GFLOPS floating point performance.

97 MATHEMATICS AND COMPUTING↗

Productive Programming of Distributed Systems with the SHAD C++ Library

High-performance computing (HPC) is often perceived as a matter of making large-scale systems (e.g., clusters) run as fast as possible, regardless the required programming effort. However, the idea of "bringing HPC to the masses" has recently emerged. Inspired by this vision, we have designed SHAD, the Scalable High-performance Algorithms and Data-structures library. SHAD is open source software, written in C++, for C++ developers. Unlike other HPC libraries for distributed systems, which rely on SPMD models, SHAD adopts a shared-memory programming abstraction, to make C++ programmers feel at home. Underneath, SHAD manages tasking and data-movements, moving the computation where data resides and taking advantage of asynchrony to tolerate network latency. At the bottom of his stack, SHAD can interface with multiple runtime systems: this not only improves developer’s productivity, by hiding the complexity of such software and of the underlying hardware, but also greatly enhance code portability. Thanks to its abstraction layers, SHAD can indeed target different systems, ranging from laptops to HPC clusters, without any need for modifying the user-level code. We have prototyped and open-sourced the implementation of (a subset of) the C++ standard library (STL) targeting multi-node HPC clusters. Our work allows plain STL-based C++ code to scale on HPC systems, with no need for rewriting the code to exploit the complex hardware. SHAD is available under Apache v2 License at https://github.com/pnnl/SHAD. In this paper we overview the design of the SHAD library, depicting its main components: runtime systems abstractions for tasking; parallel and distributed data-structures; STL-compliant interfaces and algorithms.

Castellana, Vito G.↗

A Novel Application of Parallel Betweenness Centrality to Power Grid Contingency Analysis

In Energy Management Systems, contingency analysis is commonly performed for identifying and mitigating potentially harmful power grid component failures. The exponentially increasing combinatorial number of failure modes imposes a significant computational burden for massive contingency analysis. It is critical to select a limited set of high-impact contingency cases within the constraint of computing power and time requirements to make it possible for real-time power system vulnerability assessment. In this paper, we present a novel application of parallel betweenness centrality to power grid contingency selection. We cross-validate the proposed method using the model and data of the western US power grid, and implement it on a Cray XMT system - a massively multithreaded architecture - leveraging its advantages for parallel execution of irregular algorithms, such as graph analysis. We achieve a speedup of 55 times (on 64 processors) compared against the single-processor version of the same code running on the Cray XMT. We also compare an OpenMP-based version of the same code running on an HP Superdome shared-memory machine. The performance of the Cray XMT code shows better scalability and resource utilization, and shorter execution time for large-scale power grids. This proposed approach has been evaluated in PNNL’s Electricity Infrastructure Operations Center (EIOC). It is expected to provide a quick and efficient solution to massive contingency selection problems to help power grid operators to identify and mitigate potential widespread cascading power grid failures in real time.

24 POWER TRANSMISSION AND DISTRIBUTION↗

CORE-BFS: Communication-Optimized REctangular-partitioned BFS Achieving 160.845 TeraTEPS on Frontier Supercomputer

Distributed Breadth-First Search (BFS) is fundamental to many large-scale graph applications, but its performance on parallel systems is often limited by high communication overhead. This paper presents CORE-BFS, an extremely scalable GPU-based BFS implementation that introduces a unique rectangular 2D partitioning-based design for Frontier supercomputer. To further improve performance, we propose four key optimizations: (1) Rectangular 2D-partition specific data formats that use two compressed row and one compressed column status array bitmaps combined with a Double Compressed Sparse Row (DCSR) format per partition, reducing memory footprint and inter-rank traffic; (2) Adaptive frontier & communication strategy that unifies top-down and bottom-up traversal on the rectangular layout, uses lazy synchronization in top-down levels, and switches variants based on frontier size to minimize communication overhead; (3) Frontier-split degree-aware update that maps frontier vertices to thread-centric, wavefront-centric, and block-centric kernels based on their degree to improve GPU utilization and memory coalescing; (4) Row-reduction pipeline that overlaps bottom-up adjacency list processing with row-wise bitmap reduction to hide inter-rank latency. Together, these techniques increase parallelism while reducing memory and communication overhead. On the Graph500 benchmark, CORE - BFS scales up to 9,248 Frontier nodes with scale-42 graphs and reaches 160.845 TTEPS, delivering a 5.42 × speedup over our previous Frontier implementation.

Yang, Haoshen [Rutgers University]↗

A scalable approach to solving dense linear algebra problems on hybrid CPU-GPU systems

Aiming to fully exploit the computing power of all CPUs and all graphics processing units (GPUs) on hybrid CPU-GPU systems to solve dense linear algebra problems, in this paper we design a class of heterogeneous tile algorithms to maximize the degree of parallelism, to minimize the communication volume, and to accommodate the heterogeneity between CPUs and GPUs. The new heterogeneous tile algorithms are executed upon our decentralized dynamic scheduling runtime system, which schedules a task graph dynamically and transfers data between compute nodes automatically. The runtime system uses a new distributed task assignment protocol to solve data dependencies between tasks without any coordination between processing units. By overlapping computation and communication through dynamic scheduling, we are able to attain scalable performance for the double-precision Cholesky factorization and QR factorization. Finally, our approach demonstrates a performance comparable to Intel MKL on shared-memory multicore systems and better performance than both vendor (e.g., Intel MKL) and open source libraries (e.g., StarPU) in the following three environments: heterogeneous clusters with GPUs, conventional clusters without GPUs, and shared-memory systems with multiple GPUs.

97 MATHEMATICS AND COMPUTING↗

On-the-fly Doppler Broadening with Probability Table Interpolation for Unresolved Resonance Region in RMC

With the rapid development of computational power and parallel algorithms, Monte Carlo method has been widely investigated and used in neutron transport simulations due to its advantages in geometry modelling and usage of continuous energy point-wise cross sections. Nowadays the road has been paved to couple Monte Carlo codes with thermal-hydraulics codes to make multi-physics calculations. One of the most important aspects of thermal-hydraulics feedback is the detailed temperature distribution, which demands temperature dependent cross sections in neutron transport simulations. The conventional pre-generated cross section libraries at each temperature point are no longer feasible for detailed temperature distributions because of tremendous memory footprint. Hence, different on-the-fly Doppler broadening methods have been proposed for resolved resonance energy region and thermal energy region, such as the interpolation, regress model fitting, multipole representation (MPR), target motion sampling (TMS) and different integration algorithms based on Doppler broadening equation. Compared to resolved resonance energy region and thermal energy region, the representation of cross sections in unresolved resonance region is slightly different. As a result of the limitation of experiment measurements, the overlapped resonances cannot be distinguished from each other and the statistic average resonance parameters are given in the evaluated nuclear data file. In early applications, infinite-dilute pointwise cross sections were obtained from average parameters and were directly used in neutron simulations. However, it was proven that these infinite-dilute pointwise cross sections cannot capture the energy self-shielding effect, thus leading to an overestimate of reaction rates and usually an underestimate of k{sub eff} owing to the increased absorption rates and increased scattering probability to energy below unresolved resonance region where the capture-to-fission ratio is typically higher. This phenomenon is of great importance especially for problems where there is a considerable flux in unresolved resonance region. Then the probability table method was put forward by Levitt to take energy self-shielding effect into account, which provides probability distribution function of cross sections in unresolved resonance region instead of specific cross section values. The probability distribution function is presented in the form of tables and the mean values are equal to the infinite dilution cross sections. To obtain the temperature dependent cross sections in unresolved energy region, there are two different methods. One is to calculate cross sections directly from average resonance parameters and their statistical distributions , which is effective and precise but time-consuming. The other is to interpolate between pre-generated probability table files, which is much more efficient. In this paper, on-the-fly Doppler broadening of cross sections in unresolved resonance regions based on probability tables interpolation was implemented in RMC codes as a supplement to on-the-fly Doppler broadening for resolved resonance energy region and thermal region, which is based on the integration of Doppler broadening equation. The results of criticality calculation and depletion calculation show that on-the-fly Doppler broadening for the whole energy region is accurate and efficient. (authors)

73 NUCLEAR PHYSICS AND RADIATION PHYSICS↗

The OVIS analysis architecture.

This report summarizes the current statistical analysis capability of OVIS and how it works in conjunction with the OVIS data readers and interpolators. It also documents how to extend these capabilities. OVIS is a tool for parallel statistical analysis of sensor data to improve system reliability. Parallelism is achieved using a distributed data model: many sensors on similar components (metaphorically sheep) insert measurements into a series of databases on computers reserved for analyzing the measurements (metaphorically shepherds). Each shepherd node then processes the sheep data stored locally and the results are aggregated across all shepherds. OVIS uses the Visualization Tool Kit (VTK) statistics algorithm class hierarchy to perform analysis of each process's data but avoids VTK's model aggregation stage which uses the Message Passing Interface (MPI); this is because if a single process in an MPI job fails, the entire job will fail. Instead, OVIS uses asynchronous database replication to aggregate statistical models. OVIS has several additional features beyond those present in VTK that, first, accommodate its particular data format and, second, improve the memory and speed of the statistical analyses. First, because many statistical algorithms are multivariate in nature and sensor data is typically univariate, interpolation of data is required to provide simultaneous observations of metrics. Note that in this report, we will refer to a single value obtained from a sensor as a measurement while a collection of multiple sensor values simultaneously present in the system is an observation. A base class for interpolation is provided that abstracts the operation of converting multiple sensor measurements into simultaneous observations. A concrete implementation is provided that performs piecewise constant temporal interpolation of multiple metrics across a single component. Secondly, because calculations may summarize data too large to fit in memory OVIS analyses batches of observations at a time and aggregates these intermediate intra-process models as it goes before storing the final model for inter-process aggregation via database replication. This reduces the memory footprint of the analysis, interpolation, and the database client and server query processing. This also interleaves processing with the disk I/O required to fetch data from the database - also improving speed. This report documents how OVIS performs analyses and how to create additional analysis components that fetch measurements from the database, perform interpolation, or perform operations on streamed observations (such as model updates or assessments). The rest of this section outlines the OVIS analysis algorithm and is followed by sections specific to each subtask. Note that we are limiting our discussion for now to the creation of a model from a set of measurements, and not including the assessment of observations using a model. The same framework can be used for assessment but that use case is not detailed in this report.

97 MATHEMATICS AND COMPUTING↗

Implementation of Distributed Memory Computing in MOSAIC to Enable Large 3D Simulations of Irradiated Concrete

The concrete biological shield (CBS) of light-water reactors protects workers and the surrounding environment by absorbing neutron and gamma irradiation emitted from the reactor core. The radiation dose increases with the CBS’s operational time and, in the long term, becomes significant enough to raise the question of irradiation effects on concrete—and particularly on the structural integrity of the CBS. Irradiation-induced damage has been identified as one of the main degradation mechanisms in the CBS. Neutron radiation causes the swelling of aggregate-forming minerals at different rates and amplitudes depending on the mineral’s nature. Silicate-bearing minerals such as quartz are particularly sensitive to neutron radiation and experience up to 17.8% volumetric expansion. Aggregates comprise several minerals with different orientations and are, therefore, subject to cracking as a result of mismatch strains. Additionally, the swelling of aggregates creates significant stresses in the surrounding cement paste matrix, which also results in crack formation. In parallel with the collection of characterization and irradiation test data, development of modeling and simulation tools for irradiated concrete is ongoing with the support of the US Department of Energy Office of Nuclear Energy’s Light Water Reactor Sustainability (LWRS) program. This effort resulted in the development and application of the fast-Fourier transform (FFT)–based code Microstructure-Oriented Scientific Analysis of Irradiated Concrete (MOSAIC) at Oak Ridge National Laboratory.

61 RADIATION PROTECTION AND DOSIMETRY↗

True Load Balancing for Matricized Tensor Times Khatri-Rao Product

MTTKRP is the bottleneck operation in algorithms used to compute the CP tensor decomposition. For sparse tensors, utilizing the compressed sparse fibers (CSF) storage format and the CSF-oriented MTTKRP algorithms is important for both memory and computational efficiency on distributed-memory architectures. Existing intelligent tensor partitioning models assume the computational cost of MTTKRP to be proportional to the total number of nonzeros in the tensor. However, this is not the case for the CSF-oriented MTTKRP on distributed-memory architectures. We outline two deficiencies of nonzero-based intelligent partitioning models when CSF-oriented MTTKRP operations are performed locally: failure to encode processors' computational loads and increase in total computation due to fiber fragmentation. We focus on existing fine-grain hypergraph model and propose a novel vertex weighting scheme that enables this model encode correct computational loads of processors. We also propose to augment the fine-grain model by fiber nets for reducing the increase in total computational load via minimizing fiber fragmentation. In this way, the proposed model encodes minimizing the load of the bottleneck processor. In conclusion, parallel experiments with real-world sparse tensors on up to 1024 processors prove the validity of the outlined deficiencies and demonstrate the merit of our proposed improvements in terms of parallel runtimes.

97 MATHEMATICS AND COMPUTING↗

Optimizing the Accelerated Recursive Doubling Algorithm for Block Tridiagonal Systems of Equations

The need to solve block tridiagonal systems with hundreds or thousands of right-hand sides for the same block tridiagonal matrix is common in a variety of disciplines. To meet this need, the Accelerated Recursive Doubling Algorithm was developed. After a right-hand side independent phase, the algorithm allows for the quick, online calculation of solutions for different right-hand sides. In this work, we present methods to optimize the Accelerated Recursive Doubling Algorithm in memory usage and computation time in a hybrid parallelization model. The right-hand side independent phase of the naïve implementation takes ≥ 11/3 the amount of memory required to store the tridiagonal matrix, while our implementation reduces the fraction to ≈ 5/3 . The right-hand side dependent phase of the naïve implementation takes ≥ 6 times the amount of memory required to store the right-hand side, while our implementation reduces the fraction to ≈ 3. The computation time for the independent phase is reduced to ≈ 2/3 times that of the naïve implementation, while the computation time for the dependent phase is reduced to ≈ 5/9 . With increasing numbers of shared-memory threads q on every distributed processing element, we have O(q) theoretical speedup.

97 MATHEMATICS AND COMPUTING↗