SEARCH · Engineering Papers
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.
This paper examines the applications most commonly run on the supercomputers at the Numerical Aerospace Simulation (NAS) facility. It analyzes the extent to which such applications are fundamentally oriented to vector computers, and whether or not they can be efficiently implemented on hierarchical memory machines, such as systems with cache memories and highly parallel, distributed memory systems.
A Choleski method is described and used to solve linear systems of equations that arise in large scale structural analysis. The method uses a novel variable-band storage scheme and is structured to exploit fast local memory caches while minimizing data access delays between main memory and vector registers. Several parallel implementations of this method are described for the CRAY-2 and CRAY Y-MP computers demonstrating the use of microtasking and autotasking directives. A portable parallel language, FORCE, is used for comparison with the microtasked and autotasked implementations. Results are presented comparing the matrix factorization times for three representative structural analysis problems from runs made in both dedicated and multi-user modes on both computers. CPU and wall clock timings are given for the parallel implementations and are compared to single processor timings of the same algorithm.
A Choleski method used to solve linear systems of equations that arise in large scale structural analyses is described. The method uses a novel variable-band storage scheme and is structured to exploit fast local memory caches while minimizing data access delays between main memory and vector registers. Several parallel implementations of this method are described for the CRAY-2 and CRAY Y-MP computers demonstrating the use of microtasking and autotasking directives. A portable parallel language, FORCE, is also used for two different parallel implementations, demonstrating the use of CRAY macrotasking. Results are presented comparing the matrix factorization times for three representative structural analysis problems from runs made in both dedicated and multi-user modes on both the CRAY-2 and CRAY Y-MP computers. CPU and wall clock timings are given for the various parallel methods and are compared to single processor timings of the same algorithm. Computation rates over 1 GIGAFLOP (1 billion floating point operations per second) on a four processor CRAY-2 and over 2 GIGAFLOPS on an eight processor CRAY Y-MP are demonstrated as measured by wall clock time in a dedicated environment. Reduced wall clock times for the parallel methods relative to the single processor implementation of the same Choleski algorithm are also demonstrated for runs made in multi-user mode.
Raghavan had earlier developed the software package DCSPACK which can be used for solving sparse linear systems where the coefficient matrix is symmetric and positive definite (this project was not funded by NASA but by agencies such as NSF). DSCPACK-S is the serial code and DSCPACK-P is a parallel implementation suitable for multiprocessors or networks-of-workstations with message passing using MCI. The main algorithm used is the Cholesky factorization of a sparse symmetric positive positive definite matrix A = LL(T). The code can also compute the factorization A = LDL(T). The complexity of the software arises from several factors relating to the sparsity of the matrix A. A sparse N x N matrix A has typically less that cN nonzeroes where c is a small constant. If the matrix were dense, it would have O(N2) nonzeroes. The most complicated part of such sparse Cholesky factorization relates to fill-in, i.e., zeroes in the original matrix that become nonzeroes in the factor L. An efficient implementation depends to a large extent on complex data structures and on techniques from graph theory to reduce, identify, and manage fill. DSCPACK is based on an efficient multifrontal implementation with fill-managing algorithms and implementation arising from earlier research by Raghavan and others. Sparse Cholesky factorization is typically a four step process: (1) ordering to compute a fill-reducing numbering, (2) symbolic factorization to determine the nonzero structure of L, (3) numeric factorization to compute L, and, (4) triangular solution to solve L(T)x = y and Ly = b. The first two steps are symbolic and are performed using the graph of the matrix. The numeric factorization step is of dominant cost and there are several schemes for improving performance by exploiting the nested and dense structure of groups of columns in the factor. The latter are aimed at better utilization of the cache-memory hierarchy on modem processors to prevent cache-misses and provide execution rates (operations/second) that are close to the peak rates for dense matrix computations. Currently, EPISCOPACY is being used in an application at NASA directed by J. Newman and M. James. We propose the implementation of efficient schemes for updating the LL(T) or LDL(T) factors computed in DSCPACK-S to meet the computational requirements of their project. A brief description is provided in the next section.
A processing device is provided which includes memory comprising data cache memory configured to store compressed data and metadata cache memory configured to store metadata, each portion of metadata comprising an encoding used to compress a portion of data. The processing device also includes at least one processor configured to compress portions of data and select, based on one or more utility level metrics, portions of metadata to be stored in the metadata cache memory. The at least one processor is also configured to store, in the metadata cache memory, the portions of metadata selected to be stored in the metadata cache memory, store, in the data cache memory, each portion of compressed data having a selected portion of corresponding metadata stored in the metadata cache memory. Each portion of compressed data, having the selected portion of corresponding metadata stored in the metadata cache memory, is decompressed.
In memory hierarchies, caches perform an important role in reducing average memory access latency. Minimizing cache misses can yield significant performance gains. As set-associative caches are widely used in modern architectures, capacity and conflict cache misses co-exist. These two types of cache misses require different optimization strategies. While cache misses are commonly studied using cache simulators, state-of-the-art simulators usually incur hundreds to thousands of times a program’s execution runtime. Moreover, a simulator has difficulty in simulating complex real hardware. To overcome these limitations, measurement methods are proposed to directly monitor program execution on real hardware via performance monitoring units. However, existing measurement-based tools either focus on capacity cache misses or do not distinguish capacity and conflict cache misses. In this paper, we design and implement CCProf, a lightweight measurement-based profiler that identifies conflict cache misses and associates them with program source code and data structures. CCProf incurs moderate runtime overhead that is at least an order of magnitude lower than simulators. With the evaluation on a number of representative programs, CCProf is able to guide optimizations on cache conflict misses and obtain nontrivial speedups.
A data processing apparatus includes one or more host processors with first processing units, one or more caches with second processing unit, a non-cache memory having a third processing unit and a reorder buffer operable to maintain data order during execution of a program of instructions. An instruction scheduler routes instructions to the processing units. Data coherence is maintained by control logic that blocks access to data locations in use by a selected processing unit other than the selected processing unit until data associated with the data locations are released from the reorder buffer. Data stored in the cache is written to the memory if it is already in a modified state, otherwise the state is set to the modified state. A memory controller may be used to restrict access to memory locations to be operated on.
An architectural design study was made of a multiprocessor computing system intended to meet functional and performance specifications appropriate to a manned space station application. Intermetrics, previous experience, and accumulated knowledge of the multiprocessor field is used to generate a baseline philosophy for the design of a future SUMC* multiprocessor. Interrupts are defined and the crucial questions of interrupt structure, such as processor selection and response time, are discussed. Memory hierarchy and performance is discussed extensively with particular attention to the design approach which utilizes a cache memory associated with each processor. The ability of an individual processor to approach its theoretical maximum performance is then analyzed in terms of a hit ratio. Memory management is envisioned as a virtual memory system implemented either through segmentation or paging. Addressing is discussed in terms of various register design adopted by current computers and those of advanced design.
Nonvolatile memory NVDIMMs, available as Intel Optane, are less expensive than DRAM and bring large byte-addressable storage within reach to many applications. Evaluations on graph analytics have shown promising performance only when DRAM is used as a hardware cache (Memory mode). An open question is whether graph applications can exploit Optane and DRAM directly (AppDirect mode) and achieving better-than-DRAM average bandwidth and run times. We evaluate Optane as a volatile pool on two large-scale graph applications with very different computational patterns, Grappolo and Ripples. We show that AppDirect mode can deliver better-than-DRAM performance, by allocating data structures to Optane and DRAM according to their access characteristics, resulting in higher average memory bandwidth and lower average latency. Memory mode provides DRAM-competitive performance with capacity equal to persistent memory. We demonstrate occasional 4x improvement using the latest AppDirect option and frequently observe competitive performance between Optane AppDirect Memory modes and DRAM.
As applications continue to generate multi-dimensional data at exponentially increasing rates, fast analytics to extract meaningful results is becoming extremely important. The database community has developed array databases that alleviate this problem through a series of techniques. In-situ mechanisms provide direct access to raw data in the original format---without loading and partitioning. Parallel processing scales to the largest datasets. In-memory caching reduces latency when the same data are accessed across a workload of queries. However, we are not aware of any work on distributed caching of multi-dimensional raw arrays. In this paper, we introduce a distributed framework for cost-based caching of multi-dimensional arrays in native format. Given a set of files that contain portions of an array and an online query workload, the framework computes an effective caching plan in two stages. First, the plan identifies the cells to be cached locally from each of the input files by continuously refining an evolving R-tree index. In the second stage, an optimal assignment of cells to nodes that collocates dependent cells in order to minimize the overall data transfer is determined. We design cache eviction and placement heuristic algorithms that consider the historical query workload. A thorough experimental evaluation over two real datasets in three file formats confirms the superiority -- by as much as two orders of magnitude -- of the proposed framework over existing techniques in terms of cache overhead and workload execution time.
A three-dimensional, high-order, conservative, and efficient discontinuous spectral volume (SV) method for the solutions of Maxwell's equations on unstructured grids is presented. The concept of discontinuous 2nd high-order loca1 representations to achieve conservation and high accuracy is utilized in a manner similar to the Discontinuous Galerkin (DG) method, but instead of using a Galerkin finite-element formulation, the SV method is based on a finite-volume approach to attain a simpler formulation. Conventional unstructured finite-volume methods require data reconstruction based on the least-squares formulation using neighboring cell data. Since each unknown employs a different stencil, one must repeat the least-squares inversion for every cell at each time step, or to store the inversion coefficients. In a high-order, three-dimensional computation, the former would involve impractically large CPU time, while for the latter the memory requirement becomes prohibitive. In the SV method, one starts with a relatively coarse grid of triangles or tetrahedra, called spectral volumes (SVs), and partition each SV into a number of structured subcells, called control volumes (CVs), that support a polynomial expansion of a desired degree of precision. The unknowns are cell averages over CVs. If all the SVs are partitioned in a geometrically similar manner, the reconstruction becomes universal as a weighted sum of unknowns, and only a few universal coefficients need to be stored for the surface integrals over CV faces. Since the solution is discontinuous across the SV boundaries, a Riemann solver is thus necessary to maintain conservation. In the paper, multi-parameter and symmetric SV partitions, up to quartic for triangle and cubic for tetrahedron, are first presented. The corresponding weight coefficients for CV face integrals in terms of CV cell averages for each partition are analytically determined. These discretization formulas are then applied to the integral form of the Maxwell equations. All numerical procedures for outer boundary, material interface, zonal interface, and interior SV face are unified with a single characteristic formulation. The load balancing in a massive parallel computing environment is therefore easier to achieve. A parameter is introduced in the Riemann solver to control the strength of the smoothing term. Important aspects of the data structure and its effects to communication and the optimum use of cache memory are discussed. Results will be presented for plane TE and TM waves incident on a perfectly conducting cylinder for up to fifth order of accuracy, and a plane wave incident on a perfectly conducting sphere for up to fourth order of accuracy. Comparisons are made with exact solutions for these cases.
An adaptive Kalman filter design that utilizes recursive maximum likelihood parameter identification is discussed. At the center of this design is the Kalman filter itself, which has the responsibility for attitude determination. At the same time, the identification algorithm is continually identifying the system parameters. The approach is applicable to nonlinear, as well as linear systems. This adaptive Kalman filter design has much potential for real time implementation, especially considering the fast clock speeds, cache memory and internal RAM available today. The recursive maximum likelihood algorithm is discussed in detail, with special attention directed towards its unique matrix formulation. The procedure for using the algorithm is described along with comments on how this algorithm interacts with the Kalman filter.
The design of a defect-tolerant control circuit for a set-associative cache memory is presented. The circuit maintains the stack ordering necessary for implementing the Least Recently Used (LRU) replacement algorithm. A discussion of programming techniques for bypassing defective blocks is included.
A new, high-order, conservative, and efficient method for conservation laws on unstructured grids is developed. It combines the best features of structured and unstructured grid methods to attain computational efficiency and geometric flexibility; it utilizes the concept of discontinuous and high-order local representations to achieve conservation and high accuracy; and it is based on the finite-difference formulation for simplicity. Universal reconstructions are obtained by distributing unknowns in a geometrically similar manner for all unstructured cells. Placements of the unknown and flux points with various order of accuracy are given for the line, triangular and tetrahedral elements. The data structure of the new method permits an optimum use of cache memory, resulting in further computational efficiency on modern computers. A new pointer system is developed that reduces memory requirements and simplifies programming for any order of accuracy. Numerical solutions are presented and compared with the exact solutions for wave propagation problems in both two and three dimensions to demonstrate the capability of the method. Excellent agreement has been found. The method is simpler and more efficient than previous discontinuous Galerkin and spectral volume methods for unstructured grids.
Hyrax OPeNDAP services are widely used by the Earth Science data centers in NASA, NOAA and other organizations to serve end users. In this talk, we will present some key features added in the HDF5 Hyrax OPeNDAP handler that can help data centers to better serve the HDF5netCDF-4 data products. Among these new features, we will focus on the following:1.The DAP4 support 2.The memory cache and the disk cache support that can reduce the service access time 3.The enhancement that makes the swath-like HDF5 products visualized by CF-client tools. We will also discuss the role of the HDF5 handler in-depth in the recent study of the Hyrax service in the cloud environment.
The use of complex single and multicore processors with significant cache memory, on-chip peripherals, memory controllers, and high speed input/output (IO) that integrate many of the parts of a traditional computer system is becoming more common in space applications. Such devices are often referred to as system on a chip devices (SOCs), even though the term is used somewhat inaccurately due to the lack of analog and mixed signal subcircuits. These devices are complex combinations of single- or multi-core processors with memory controllers, high-speed input/output (IO), and other peripheral structures that formerly would have been handled by off-chip resources. In the past the processors were tested for single event effects (SEE) separately, and the peripherals were often put into custom application-specific integrated circuits (ASICs) along with other resources required by the user. Performance and cost pressures have pushed commercial devices to incorporate many of the functional blocks into a single chip, an SOC. The NASA Electronic Parts and Packaging Program (NEPP) has been examining ways to perform SEE radiation hardness assurance (RHA) testing of these processor-centric SOCs to achieve reasonable understanding of their performance in space missions.
A multiprocessor system supports multiple concurrent modes of speculative execution. Speculation identification numbers (IDs) are allocated to speculative threads from a pool of available numbers. The pool is divided into domains, with each domain being assigned to a mode of speculation. Modes of speculation include TM, TLS, and rollback. Allocation of the IDs is carried out with respect to a central state table and using hardware pointers. The IDs are used for writing different versions of speculative results in different ways of a set in a cache memory.
Systems and methods of building massively parallel computing systems using low power computing complexes in accordance with embodiments of the invention are disclosed. A massively parallel computing system in accordance with one embodiment of the invention includes at least one Solid State Blade configured to communicate via a high performance network fabric. In addition, each Solid State Blade includes a processor configured to communicate with a plurality of low power computing complexes interconnected by a router, and each low power computing complex includes at least one general processing core, an accelerator, an I/O interface, and cache memory and is configured to communicate with non-volatile solid state memory.