Search NASA⌕ Search

SEARCH · Search NASA

Results for “memory access optimization”

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 55 records · Page 3

An Integrated Framework for Memory-Centric Analysis: From Trace Collection to Co-Design

The memory wall phenomenon—where advances in processor performance significantly outpace those in memory subsystems—poses a fundamental challenge for contemporary computing systems. In memory-bound applications, memory subsystem behavior dominates performance, yet existing analysis approaches present significant limitations: detailed microarchitectural simulators require days to weeks to simulate modest workloads; hardware performance counters provide only aggregate statistics that obscure temporal and spatial access patterns; and scaled simulation approaches face challenges in capturing certain behaviors that emerge at larger scales. These limitations reflect a processor-centric design philosophy increasingly misaligned with memory-bound workloads where detailed understanding of memory access patterns, cache hierarchy interactions, and contention is critical for effective optimization. This paper presents an integrated framework for memory-centric analysis that enables effective hardware-software co-design. We describe practical trace collection techniques, including hardware-assisted processor tracing with minimal overhead and portable software-based instrumentation with statistical sampling. We present multi-perspective analysis methods that examine memory behavior from temporal, sequential, spatial, and relational viewpoints, revealing distinct optimization opportunities invisible in aggregate metrics. We detail an architectural modeling framework that uses sampled traces with temporal interpolation and confidence-based filtering to evaluate cache and memory configurations. Evaluation on representative benchmarks demonstrates that this framework achieves practical accuracy (L2 cache errors of 2.64\%, confidence-filtered L3 errors of 9.92\%, bandwidth errors of 7.33\%) while providing substantial speedup (26.8×) over cycle-accurate simulation, enabling rapid design space exploration. We demonstrate how this integrated framework enables systematic identification of both hardware optimizations (memory controller tuning, bank partitioning, NUMA configuration) and software optimizations (data layout restructuring, prefetching strategies, memory-aware scheduling). Through this comprehensive treatment of the memory-centric analysis pipeline—from trace collection through architectural modeling to co-design application—we provide researchers and practitioners with practical techniques for addressing memory bottlenecks in contemporary computing systems.

Gajaria, Dhruv Mayur↗

Static multiresolution grids with inline hierarchy information for cosmic ray propagation

For numerical simulations of cosmic-ray propagation fast access to static magnetic field data is required. We present a data structure for multiresolution vector grids which is optimized for fast access, low overhead and shared memory use. The hierarchy information is encoded into the grid itself, reducing the memory overhead. Benchmarks show that in certain scenarios the differences in deflections introduced by sampling the magnetic field model can be significantly reduced when using the multiresolution approach.

79 ASTRONOMY AND ASTROPHYSICS↗

Interesting Characteristics of AMD Barcelona Floating Point Execution

In almost all modern scientific applications, developers achieve the greatest performance gains by tuning algorithms, communication systems, and memory access patterns, while leaving low level instruction optimizations to the compiler. Given the increasingly varied and complicated x86 architectures, the value of these optimizations is unclear, and, due to time and complexity constraints, it is difficult for many programmers to experiment with them. In this report we explore the potential gains of these 'last mile' optimization efforts on an AMD Barcelona processor, providing readers with relevant information so that they can decide whether investment in the presented optimizations is worthwhile.

97 MATHEMATICS AND COMPUTING↗

Critical Points Based Register-Concurrency Autotuning for GPUs

The unprecedented prevalence of GPGPU is largely attributed to its abundant on-chip register resources, which allow massively concurrent threads and extremely fast context switch. However, due to internal memory capacity constraints, there is a tradeoff between the per-thread register usage and the overall concurrency. This becomes a design problem in terms of performance tuning, since the performance “sweet spot” which can be significantly affected by these two factors is generally unknown beforehand. In this paper, we propose an effective autotuning solution to quickly and efficiently select the optimal number of registers perthread for delivering the best GPU performance. Experiments on three generations of GPUs (Nvidia Fermi, Kepler and Maxwell) demonstrate that our simple strategy can achieve an average of 10% performance improvement while a max of 50% over the original version without modifying the user program. Additionally, to reduce local cache misses due to register spilling and further improve performance, we explore three optimization schemes (i.e. bypass L1 for global memory access, enlarge local L1 cache and spill into shared memory) and discuss their impact on performance on a Kepler GPU.

auto-tuning↗

Challenges of Monte Carlo Transport

These are slides from a presentation for Parallel Summer School at Los Alamos National Laboratory. Solving discretized partial differential equations (PDEs) of interest can require a large number of computations. We can identify concurrency to allow parallel solution of discrete PDEs. Simulated particles histories can be used to solve the Boltzmann transport equation. Particle histories are independent in neutral particle transport, making them amenable to parallel computation. Physical parameters and method type determine the data dependencies of particle histories. Data requirements shape parallel algorithms for Monte Carlo. Then, Parallel Computational Physics and Parallel Monte Carlo are discussed and, finally, the results are given. The mesh passing method greatly simplifies the IMC implementation and allows simple load-balancing. Using MPI windows and passive, one-sided RMA further simplifies the implementation by removing target synchronization. The author is very interested in implementations of PGAS that may allow further optimization for one-sided, read-only memory access (e.g. Open SHMEM). The MPICH_RMA_OVER_DMAPP option and library is required to make one-sided messaging scale on Trinitite - Moonlight scales poorly. Interconnect specific libraries or functions are likely necessary to ensure performance. BRANSON has been used to directly compare the current standard method to a proposed method on idealized problems. The mesh passing algorithm performs well on problems that are designed to show the scalability of the particle passing method. BRANSON can now run load-imbalanced, dynamic problems. Potential avenues of improvement in the mesh passing algorithm will be implemented and explored. A suite of test problems that stress DD methods will elucidate a possible path forward for production codes.

73 NUCLEAR PHYSICS AND RADIATION PHYSICS↗

Switching current reduction in magnetoresistive random access memories

We present an approach for minimizing the critical current for the magnetization switching in magnetic tunnel junctions by optimizing the spatial distribution of the current density. We show that such a minimization is possible because critical current is determined by the condition of making one of the magnetization eigenstates grow in time. The excitation of the eigenstates is enhanced when the spatial distributions of the eigenstates and current density overlap. Critical current can be viewed as a functional of the current density spatial distribution and it can be minimized by optimizing this distribution. Such an optimization results in a major reduction of the critical current and increase of the switching efficiency, viz. the ratio between the energy barrier and critical current. The minimized critical current increases approximately linearly with the magnetic tunnel junction size, which is much slower than critical current for the case of a uniform current density. The optimized efficiency can be approximately a constant with respect to the magnetic tunnel junction size, which is much higher than the efficiency for the uniform current density. Additional optimization can be achieved by spatially modulating the material parameters, e.g., the saturation magnetization. The presented approach and obtained scaling of the critical current and efficiency offers opportunities for the magnetic tunnel junction optimization.

36 MATERIALS SCIENCE↗

MemLeap GPU Memory Profiler

MemLeap is a GPU Memory Profiler built atop Nvbit (NVidia Binary Instrumentation Tool) framework. It analyzes memory accesses to memory objects in GPU memory, including Unified Memory, to optimize memory usage and reveal performance bottlenecks.

Peng, Bo↗

Performance of asynchronous optimized Schwarz with one-sided communication

In asynchronous iterative methods on distributed-memory computers, processes update their local solutions using data from other processes without an implicit or explicit global synchronization that corresponds to advancing the global iteration counter. Here in this work, we test the asynchronous optimized Schwarz domain-decomposition iterative method using various one-sided (remote direct memory access) communication schemes with passive target completion. The results show that when one-sided communication is well-supported, the asynchronous version of optimized Schwarz can outperform the synchronous version even for perfectly balanced partitionings of the problem on a supercomputer with uniform nodes.

97 MATHEMATICS AND COMPUTING↗

Optimizing TLB entries for mixed page size storage in contiguous memory

A system and method for accessing memory are provided. The system comprises a lookup buffer for storing one or more page table entries, wherein each of the one or more page table entries comprises at least a virtual page number and a physical page number; a logic circuit for receiving a virtual address from said processor, said logic circuit for matching the virtual address to the virtual page number in one of the page table entries to select the physical page number in the same page table entry, said page table entry having one or more bits set to exclude a memory range from a page.

97 MATHEMATICS AND COMPUTING↗

Optimizing TLB entries for mixed page size storage in contiguous memory

A system and method for accessing memory are provided. The system comprises a lookup buffer for storing one or more page table entries, wherein each of the one or more page table entries comprises at least a virtual page number and a physical page number; a logic circuit for receiving a virtual address from said processor, said logic circuit for matching the virtual address to the virtual page number in one of the page table entries to select the physical page number in the same page table entry, said page table entry having one or more bits set to exclude a memory range from a page.

Chen, Dong↗

Improving the Performance of DGEMM with MoA and Cache-Blocking: Preprint

The goal of this paper is to demonstrate performance enhancements of the high performance dense linear algebra matrix-matrix multiply DGEMM kernel, widely implemented by vendors in the basic linear algebra subroutine BLAS library. The mathematics of arrays (MoA) paradigm due to Mullin (1988) results in contiguous memory accesses in combination with Church-Rosser complete language constructs optimized for target processor architectures [3]. Our performance studies demonstrate that the MoA implementation of DGEMM combined with optimal cache-blocking strategies results in at least a 25% performance gain on both Intel Xeon Skylake and IBM Power-9 processors over the vendor supplied Intel MKL and IBM ESSL basic linear algebra libraries. Results are presented for the NREL Eagle and ORNL Summit supercomputers.

cache-blocking↗

Performance Optimization and Auto-Tuning

In the broader computational research community, one subject of recent research is the problem of adapting algorithms to make effective use of multi- and many-core processors. Effective use of these architectures, which have complex memory hierarchies with many layers of cache, typically involves a careful examination of how an algorithm moves data through the memory hierarchy. Unfortunately, there is often a non-obvious relationship between algorithmic parameters like blocking strategies, and their impact on memory utilization, and, in turn, the relationship with runtime performance. Auto-tuning is an empirical method used to discover optimal values for tunable algorithmic parameters under such circumstances. The challenge is compounded by the fact that the settings that produce the best performance for a given problem and a given platform may not be the best for a different problem on the same platform, or the same problem on a different platform. The high performance visualization research community has begun to explore and adapt the principles of auto-tuning for the purpose of optimizing codes on modern multi- and many-core processors. This report focuses on how performance optimization studies reveal a dramatic variation in performance for two fundamental visualization algorithms: one based on a stencil operation having structured, uniform memory access, and the other is ray casting volume rendering, which uses unstructured memory access patterns. The two case studies highlighted in this report show the extra effort required to optimize such codes by adjusting the tunable algorithmic parameters can return substantial gains in performance. Additionally, these case studies also explore the potential impact of and the interaction between algorithmic optimizations and tunable algorithmic parameters, along with the potential performance gains resulting from leveraging architecture-specific features.

97 MATHEMATICS AND COMPUTING↗

Exploration of Optimization Options for Increasing Performance of a GPU Implementation of a Three-dimensional Bilateral Filter

This report explores using GPUs as a platform for performing high performance medical image data processing, specifically smoothing using a 3D bilateral filter, which performs anisotropic, edge-preserving smoothing. The algorithm consists of a running a specialized 3D convolution kernel over a source volume to produce an output volume. Overall, our objective is to understand what algorithmic design choices and configuration options lead to optimal performance of this algorithm on the GPU. We explore the performance impact of using different memory access patterns, of using different types of device/on-chip memories, of using strictly aligned and unaligned memory, and of varying the size/shape of thread blocks. Our results reveal optimal configuration parameters for our algorithm when executed sample 3D medical data set, and show performance gains ranging from 30x to over 200x as compared to a single-threaded CPU implementation.

97 MATHEMATICS AND COMPUTING↗

System, methods and apparatus for program optimization for multi-threaded processor architectures

Methods, apparatus and computer software product for source code optimization are provided. In an exemplary embodiment, a first custom computing apparatus is used to optimize the execution of source code on a second computing apparatus. In this embodiment, the first custom computing apparatus contains a memory, a storage medium and at least one processor with at least one multi-stage execution unit. The second computing apparatus contains at least two multi-stage execution units that allow for parallel execution of tasks. The first custom computing apparatus optimizes the code for parallelism, locality of operations and contiguity of memory accesses on the second computing apparatus. This Abstract is provided for the sole purpose of complying with the Abstract requirement rules. This Abstract is submitted with the explicit understanding that it will not be used to interpret or to limit the scope or the meaning of the claims.

42 ENGINEERING↗

The Art of War: Beyond Memory-one Strategies in Population Games

We show that the history of play in a population game contains exploitable information that can be successfully used by sophisticated strategies to defeat memory-one opponents, including zero determinant strategies. The history allows a player to label opponents by their strategies, enabling a player to determine the population distribution and to act differentially based on the opponent’s strategy in each pairwise interaction. For the Prisoner’s Dilemma, these advantages lead to the natural formation of cooperative coalitions among similarly behaving players and eventually to unilateral defection against opposing player types. We show analytically and empirically that optimal play in population games depends strongly on the population distribution. For example, the optimal strategy for a minority player type against a resident TFT population is ALLC, while for a majority player type the optimal strategy versus TFT players is ALLD. Such behaviors are not accessible to memory-one strategies. Drawing inspiration from Sun Tzu’s the Art of War, we implemented a non-memory-one strategy for population games based on techniques from machine learning and statistical inference that can exploit the history of play in this manner. Via simulation we find that this strategy is essentially uninvadable and can successfully invade (significantly more likely than a neutral mutant) essentially all known memory-one strategies for the Prisoner’s Dilemma, including ALLC (always cooperate), ALLD (always defect), tit-for-tat (TFT), winstay-lose-shift (WSLS), and zero determinant (ZD) strategies, including extortionate and generous strategies.

60 APPLIED LIFE SCIENCES↗

In Situ Indexing and Query Processing of AMR Data

We have developed the first, holistic, Adaptive Mesh Refinement (AMR) capable, in situ query framework to ensure the tractability of the current- and next-generation AMR simulation analysis on the path to extreme scale. We have addressed key issues pertaining to in situ analysis, memory hierarchy, generalized parallel indexing, index-compression, and real-time access pattern detection. We have published the following research papers in high profile peer reviewed computer science conferences: (a) Parallel In Situ Detection of Connected Components in Adaptive Mesh Refinement Data; (b) AMR-aware In Situ Indexing and Scalable Querying; (c) Pattern-driven dynamic data layout optimization; (d) In situ Storage Layout Optimization for AMR Spatio-temporal Read Accesses; (e) AMRZone: A Runtime AMR Data Sharing Framework for Scientific Applications; (f) Exploring Memory Hierarchy and Network Topology for Runtime AMR Data Sharing Across Scientific Applications.

96 KNOWLEDGE MANAGEMENT AND PRESERVATION↗

Multi-core and Many-core Shared-memory Parallel Raycasting Volume Rendering Optimization and Tuning

Given the computing industry trend of increasing processing capacity by adding more cores to a chip, the focus of this work is tuning the performance of a staple visualization algorithm, raycasting volume rendering, for shared-memory parallelism on multi-core CPUs and many-core GPUs. Our approach is to vary tunable algorithmic settings, along with known algorithmic optimizations and two different memory layouts, and measure performance in terms of absolute runtime and L2 memory cache misses. Our results indicate there is a wide variation in runtime performance on all platforms, as much as 254% for the tunable parameters we test on multi-core CPUs and 265% on many-core GPUs, and the optimal configurations vary across platforms, often in a non-obvious way. For example, our results indicate the optimal configurations on the GPU occur at a crossover point between those that maintain good cache utilization and those that saturate computational throughput. This result is likely to be extremely difficult to predict with an empirical performance model for this particular algorithm because it has an unstructured memory access pattern that varies locally for individual rays and globally for the selected viewpoint. Our results also show that optimal parameters on modern architectures are markedly different from those in previous studies run on older architectures. And, given the dramatic performance variation across platforms for both optimal algorithm settings and performance results, there is a clear benefit for production visualization and analysis codes to adopt a strategy for performance optimization through auto-tuning. These benefits will likely become more pronounced in the future as the number of cores per chip and the cost of moving data through the memory hierarchy both increase.

97 MATHEMATICS AND COMPUTING↗

Multi-core and many-core shared-memory parallel raycasting volume rendering optimization and tuning

Given the computing industry trend of increasing processing capacity by adding more cores to a chip, the focus of this work is tuning the performance of a staple visualization algorithm, raycasting volume rendering, for shared-memory parallelism on multi-core CPUs and many-core GPUs. Our approach is to vary tunable algorithmic settings, along with known algorithmic optimizations and two different memory layouts, and measure performance in terms of absolute runtime and L2 memory cache misses. Our results indicate there is a wide variation in runtime performance on all platforms, as much as 254% for the tunable parameters we test on multi-core CPUs and 265% on many-core GPUs, and the optimal configurations vary across platforms, often in a non-obvious way. For example, our results indicate the optimal configurations on the GPU occur at a crossover point between those that maintain good cache utilization and those that saturate computational throughput. This result is likely to be extremely difficult to predict with an empirical performance model for this particular algorithm because it has an unstructured memory access pattern that varies locally for individual rays and globally for the selected viewpoint. Our results also show that optimal parameters on modern architectures are markedly different from those in previous studies run on older architectures. In addition, given the dramatic performance variation across platforms for both optimal algorithm settings and performance results, there is a clear benefit for production visualization and analysis codes to adopt a strategy for performance optimization through auto-tuning. These benefits will likely become more pronounced in the future as the number of cores per chip and the cost of moving data through the memory hierarchy both increase.

97 MATHEMATICS AND COMPUTING↗