Extending an asynchronous runtime system for high throughput applications: A case study
Not Available
SEARCH · Search NASA
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.
Not Available
Not Available
Prior to every ion implantation experiment a simulation of the ion range and other relevant parameters is performed using Monte-Carlo based codes. Although increasing computing power has improved the speed of these calculations, the demands on Monte-Carlo codes are also increasing, requiring evaluation of the optimal number of simulations while ensuring accuracy within threshold bounds. We evaluate the “Stopping and Range of Ions in Matter” (SRIM) code due to its widespread usage. We show how dividing simulations into multiple parallel simulations with different random seeds can lead to calculation speedup and find lower bounds for the required number of ion traces simulated based on an exemplar system of a Ga focused ion beam and a high energy C beam as used in high linear energy transfer testing. Here our results indicate simulations can yield results within the underlying data accuracy of SRIM at 10X and 100X shorter simulation time than the SRIM default values.
Abstract A leading approach to algorithm design aims to minimize the number of operations in an algorithm’s compilation. One intuitively expects that reducing the number of operations may decrease the chance of errors. This paradigm is particularly prevalent in quantum computing, where gates are hard to implement and noise rapidly decreases a quantum computer’s potential to outperform classical computers. Here, we find that minimizing the number of operations in a quantum algorithm can be counterproductive, leading to a noise sensitivity that induces errors when running the algorithm in non-ideal conditions. To show this, we develop a framework to characterize the resilience of an algorithm to perturbative noises (including coherent errors, dephasing, and depolarizing noise). Some compilations of an algorithm can be resilient against certain noise sources while being unstable against other noises. We condense these results into a tradeoff relation between an algorithm’s number of operations and its noise resilience. We also show how this framework can be leveraged to identify compilations of an algorithm that are better suited to withstand certain noises.
Classical optimization problems can be solved by adiabatically preparing the ground state of a quantum Hamiltonian that encodes the problem. The performance of this approach is determined by the smallest gap encountered during the evolution. Here, we consider the maximum independent set problem, which can be efficiently encoded in the Hamiltonian describing a Rydberg atom array. We present a general construction of instances of the problem for which the minimum gap decays superexponentially with system size, implying a superexponentially large time to solution via adiabatic evolution. The small gap arises from locally independent choices which cause the system to initially evolve and localize into a configuration far from the solution in terms of Hamming distance. We investigate remedies to this problem. Specifically, we show that quantum quenches in these models can exhibit signatures of quantum many-body scars, which in turn, can circumvent the superexponential gaps. By quenching from a suboptimal configuration, states with a larger ground-state overlap can be prepared, illustrating the utility of quantum quenches as an algorithmic tool. Published by the American Physical Society 2024
Not Available
Explore the source record for details and available documents.
Scientific workflows have become ubiquitous across scientific fields, and their execution methods and systems continue to be the subject of research and development. Most experimental evaluations of these workflows rely on workflow instances, which can be either real-world or synthetic, to ensure relevance to current application domains or explore hypothetical/future scenarios. The WfCommons project addresses this need by providing data and tools to support such evaluations. In this paper, we present an overview of WfCommons and describe two recent developments. Firstly, we introduce a workflow execution "tracer" for Nextflow, which significantly enhances the set of real-world instances available in WfCommons. Secondly, we describe a workflow instance "translator" that enables the execution of any real-world or synthetic WfCommons workflow instance using Dask. Our contributions aim to provide researchers and practitioners with more comprehensive resources for evaluating scientific workflows.
Abstract not provided.
Abstract not provided
Abstract not provided.
Explore the source record for details and available documents.
Abstract not provided.
Abstract not provided
The recent development of deep learning has been mostly focusing on Euclidean data, such as images, videos, audios, etc. However, most real-world information and relation are often expressed as graphs. To efficiently learn from graph data, graph convolutional networks (GCNs) emerge as a promising approach, showing advantages in several practical applications such as social network analysis, knowledge discovery, 3D modeling, motion capturing, etc. Real-world graphs are usually extremely large and imbalanced, posting significant performance demand and design challenges on the hardware dedicated for GCN inference. In this paper, we propose an architecture design called UW-GCN to accelerate graph convolutional network inference. To tackle the major performance bottleneck from workload imbalance, we propose dynamic neighborhood stealing and remote chunk shuffling techniques, relying on hardware flexibility to achieve hardware auto-tuning under negligible area or delay overhead. Specifically, UW-GCN is able to smartly profile the sparse graph pattern while continuously adjusting the workload distribution via routing reconfiguration among parallel processing elements (PEs). The ideal configuration is then reused in the remaining iterations. To the best of our knowledge, this is the first accelerator design particularly for GCN and the first work relying on hardware auto-tuning, which is normally based on software, to achieve near-optimal workload balance in processing sparse structures.
Abstract not provided
Abstract not provided