SEARCH · Search NASA

Results for “runtime performance”

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.

SourceResearch connectionFrom yearThrough year

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

At least 595 records · Page 33

All sources NASA NTRS DOE OSTI

DOE OSTI2017

Scalable Failure Masking for Stencil Computations using Ghost Region Expansion and Cell to Rank Remapping

In order to achieve exascale systems, application resilience needs to be addressed. Some programming models, such as task-DAG (directed acyclic graphs) architectures, currently embed resilience features whereas traditional SPMD (single program, multiple data) and message-passing models do not. Since a large part of the community's code base follows the latter models, it is still required to take advantage of application characteristics to minimize the overheads of fault tolerance. To that end, this paper explores how recovering from hard process/node failures in a local manner is a natural approach for certain applications to obtain resilience at lower costs in faulty environments. In particular, this paper targets enabling online, semitransparent local recovery for stencil computations on current leadership-class systems as well as presents programming support and scalable runtime mechanisms. Also described and demonstrated in this paper is the effect of failure masking, which allows the effective reduction of impact on total time to solution due to multiple failures. Furthermore, we discuss, implement, and evaluate ghost region expansion and cell-to-rank remapping to increase the probability of failure masking. To conclude, this paper shows the integration of all aforementioned mechanisms with the S3D combustion simulation through an experimental demonstration (using the Titan system) of the ability to tolerate high failure rates (i.e., node failures every five seconds) with low overhead while sustaining performance at large scales. In addition, this demonstration also displays the failure masking probability increase resulting from the combination of both ghost region expansion and cell-to-rank remapping.

97 MATHEMATICS AND COMPUTING↗

DOE OSTI2020

Efficient Data Management in Neutron Scattering Data Reduction Workflows at ORNL

Oak Ridge National Laboratory (ORNL) experimental neutron science facilities produce 1.2 TB a day of raw event-based data that is stored using the standard metadata-rich NeXus schema built on top of the HDF5 file format. Performance of several data reduction workflows is largely determined by the amount of time spent on the loading and processing algorithms in Mantid, an open-source data analysis framework used across several neutron sciences facilities around the world. The present work introduces new data management algorithms to address identified input output (I/O) bottlenecks on Mantid. First, we introduce an in-memory binary-tree metadata index that resemble NeXus data access patterns to provide a scalable search and extraction mechanism. Second, data encapsulation in Mantid algorithms is optimally redesigned to reduce the total compute and memory runtime footprint associated with metadata I/O reconstruction tasks. Results from this work show speed ups in wall-clock time on ORNL data reduction workflows, ranging from 11% to 30% depending on the complexity of the targeted instrument-specific data. Nevertheless, we highlight the need for more research to address reduction challenges as experimental data volumes increase.

Godoy, William↗

DOE OSTI2015

A Case for Application Oblivious Energy-Efficient MPI Runtime

Power has become the major impediment in designing large scale high-end systems. Message Passing Interface (MPI) is the {\em de facto} communication interface used as the back-end for designing applications, programming models and runtime for these systems. Slack --- the time spent by an MPI process in a single MPI call --- provides a potential for energy and power savings, if an appropriate power reduction technique such as core-idling/Dynamic Voltage and Frequency Scaling (DVFS) can be applied without perturbing application's execution time. Existing techniques that exploit slack for power savings assume that application behavior repeats across iterations/executions. However, an increasing use of adaptive, data-dependent workloads combined with system factors (OS noise, congestion) makes this assumption invalid. This paper proposes and implements Energy Aware MPI (EAM) --- an application-oblivious energy-efficient MPI runtime. EAM uses a combination of communication models of common MPI primitives (point-to-point, collective, progress, blocking/non-blocking) and an online observation of slack for maximizing energy efficiency. Each power lever incurs time overhead, which must be amortized over slack to minimize degradation. When predicted communication time exceeds a lever overhead, the lever is used {\em as soon as possible} --- to maximize energy efficiency. When mis-prediction occurs, the lever(s) are used automatically at specific intervals for amortization. We implement EAM using MVAPICH2 and evaluate it on ten applications using up to 4096 processes. Our performance evaluation on an InfiniBand cluster indicates that EAM can reduce energy consumption by 5--41\% in comparison to the default approach, with negligible (less than 4\% in all cases) performance loss.

Venkatesh, Akshay↗

DOE OSTI2026

ORCHA: A performance portability system for extreme heterogeneity

Heterogeneity is the prevalent trend in the rapidly evolving high-performance computing (HPC) landscape in both hardware and application software. The diversity in hardware platforms, currently comprising various accelerators and a future possibility of specializable chiplets, poses a significant challenge for scientific software developers aiming to harness optimal performance across different computing platforms while maintaining the quality of solutions when their applications are simultaneously growing more complex. Code synthesis and code generation can provide mechanisms to mitigate this challenge. We have developed a divide and conquer approach where different aspects of performance are handled by different stand-alone tools that are interfaced with the application through generated code. This portability system, ORCHA, enables users to configure and orchestrate their computations among available resources on a platform by specifying a high-level recipe, thereby permitting a many-to-many paradigm where each recipe results in a different variant of the application. The core design goal is to let users decide the application’s hardware mapping and orchestration by editing only the high-level recipe—without modifying the maintained source code or binding the application to a particular runtime system. Tools in ORCHA distribution are: CG-Kit for translating the recipe into an execution graph; Milhoja to execute the graph by orchestrating data and task movement among hardware resources; and Macroprocessor that enables users to define their own code-shorthand for higher composability and easier management of code variants. Additionally, the design of ORCHA permits tools to work in a plug-and-play mode where the application can build and run without CG-Kit and Milhoja, and either tool can be swapped out for other tools with similar capabilities by modifying the code generation portion of ORCHA. In this paper, we describe the design of ORCHA and the role that code-generation plays in isolating applications from tools. We demonstrate the breadth of configurations ORCHA enables with a case study in which an application configuration is realized on three distinct hardware mappings—a GPU-centric, a CPU/GPU balanced, and a CPU/GPU concurrent layouts by using different recipes.

Lee, Youngjun↗

DOE OSTI2014

Poster — Thur Eve — 69: Computational Study of DVH-guided Cancer Treatment Planning Optimization Methods

Purpose: To compare methods to incorporate the Dose Volume Histogram (DVH) curves into the treatment planning optimization. Method: The performance of three methods, namely, the conventional Mixed Integer Programming (MIP) model, a convex moment-based constrained optimization approach, and an unconstrained convex moment-based penalty approach, is compared using anonymized data of a prostate cancer patient. Three plans we generated using the corresponding optimization models. Four Organs at Risk (OARs) and one Tumor were involved in the treatment planning. The OARs and Tumor were discretized into total of 50,221 voxels. The number of beamlets was 943. We used commercially available optimization software Gurobi and Matlab to solve the models. Plan comparison was done by recording the model runtime followed by visual inspection of the resulting dose volume histograms. Conclusion: We demonstrate the effectiveness of the moment-based approaches to replicate the set of prescribed DVH curves. The unconstrained convex moment-based penalty approach is concluded to have the greatest potential to reduce the computational effort and holds a promise of substantial computational speed up.

60 APPLIED LIFE SCIENCES↗

DOE OSTI2015

In-Situ Visualization Experiments with ParaView Cinema in RAGE

A previous paper described some numerical experiments performed using the ParaView/Catalyst in-situ visualization infrastructure deployed in the Los Alamos RAGE radiation-hydrodynamics code to produce images from a running large scale 3D ICF simulation. One challenge of the in-situ approach apparent in these experiments was the difficulty of choosing parameters likes isosurface values for the visualizations to be produced from the running simulation without the benefit of prior knowledge of the simulation results and the resultant cost of recomputing in-situ generated images when parameters are chosen suboptimally. A proposed method of addressing this difficulty is to simply render multiple images at runtime with a range of possible parameter values to produce a large database of images and to provide the user with a tool for managing the resulting database of imagery. Recently, ParaView/Catalyst has been extended to include such a capability via the so-called Cinema framework. Here I describe some initial experiments with the first delivery of Cinema and make some recommendations for future extensions of Cinema’s capabilities.

in-situ visualization↗

DOE OSTI2023

Linear-depth quantum circuits for loading Fourier approximations of arbitrary functions

Abstract The ability to efficiently load functions on quantum computers with high fidelity is essential for many quantum algorithms, including those for solving partial differential equations and Monte Carlo estimation. In this work, we introduce the Fourier series loader (FSL) method for preparing quantum states that exactly encode multi-dimensional Fourier series using linear-depth quantum circuits. Specifically, the FSL method prepares a (Dn)-qubit state encoding the 2 Dn -point uniform discretization of aD-dimensional function specified by aD-dimensional Fourier series. A free parameter,m, which must be less thann, determines the number of Fourier coefficients, 2 D ( m + 1 ) , used to represent the function. The FSL method uses a quantum circuit of depth at most 2 ( n − 2 ) + ⌈ log 2 ( n − m ) ⌉ + 2 D ( m + 1 ) + 2 − 2 D ( m + 1 ) , which is linear in the number of Fourier coefficients, and linear in the number of qubits (Dn) despite the fact that the loaded function’s discretization is over exponentially many (2 Dn ) points. The FSL circuit consists of at most D n + 2 D ( m + 1 ) + 1 − 1 single-qubit and D n ( n + 1 ) / 2 + 2 D ( m + 1 ) + 1 − 3 D ( m + 1 ) − 2 two-qubit gates; we present a classical compilation algorithm with runtime O ( 2 3 D ( m + 1 ) ) to determine the FSL circuit for a given Fourier series. The FSL method allows for the highly accurate loading of complex-valued functions that are well-approximated by a Fourier series with finitely many terms. We report results from noiseless quantum circuit simulations, illustrating the capability of the FSL method to load various continuous 1D functions, and a discontinuous 1D function, on 20 qubits with infidelities of less than 10 −6 and 10 −3 , respectively. We also demonstrate the practicality of the FSL method for near-term quantum computers by presenting experiments performed on the Quantinuum H1-1 and H1-2 trapped-ion quantum computers: we loaded a complex-valued function on 3 qubits with a fidelity of over 95 % , as well as various 1D real-valued functions on up to 6 qubits with classical fidelities ≈99%, and a 2D function on 10 qubits with a classical fidelity ≈94%.

Physics↗

DOE OSTI2025

Toward Mixed Analog-Digital Quantum Signal Processing: Quantum AD/DA Conversion and the Fourier Transform

Signal processing stands as a pillar of classical computation and modern information technology, applicable to both analog and digital signals. Recently, advancements in quantum information science have suggested that quantum signal processing (QSP) can enable more powerful signal processing capabilities. However, the developments in QSP have primarily leveraged digital quantum resources, such as discrete-variable (DV) systems like qubits, rather than analog quantum resources, such as continuous-variable (CV) systems like quantum oscillators. Consequently, there remains a gap in understanding how signal processing can be performed on hybrid CV-DV quantum computers. Here we address this gap by developing a new paradigm of mixed analog-digital QSP. We demonstrate the utility of this paradigm by showcasing how it naturally enables analog-digital conversion of quantum signals—specifically, the transfer of states between DV and CV quantum systems. We then show that such quantum analog-digital conversion enables new implementations of quantum algorithms on CV-DV hardware. This is exemplified by realizing the quantum Fourier transform of a state encoded on qubits via the free-evolution of a quantum oscillator, albeit with a runtime exponential in the number of qubits due to information theoretic arguments. Collectively, this work marks a significant step forward in hybrid CV-DV quantum computation, providing a foundation for scalable analog-digital signal processing on quantum processors.

42 ENGINEERING↗

DOE OSTI2015

Simplified Predictive Models for CO 2 Sequestration Performance Assessment: Research Topical Report on Task #4 - Reduced-Order Method (ROM) Based Models

Reduced-order models provide a means for greatly accelerating the detailed simulations that will be required to manage CO 2 storage operations. In this work, we investigate the use of one such method, POD-TPWL, which has previously been shown to be effective in oil reservoir simulation problems. This method combines trajectory piecewise linearization (TPWL), in which the solution to a new (test) problem is represented through a linearization around the solution to a previously-simulated (training) problem, with proper orthogonal decomposition (POD), which enables solution states to be expressed in terms of a relatively small number of parameters. We describe the application of POD-TPWL for CO 2 -water systems simulated using a compositional procedure. Stanford’s Automatic Differentiation-based General Purpose Research Simulator (AD-GPRS) performs the full-order training simulations and provides the output (derivative matrices and system states) required by the POD-TPWL method. A new POD-TPWL capability introduced in this work is the use of horizontal injection wells that operate under rate (rather than bottom-hole pressure) control. Simulation results are presented for CO 2 injection into a synthetic aquifer and into a simplified model of the Mount Simon formation. Test cases involve the use of time-varying well controls that differ from those used in training runs. Results of reasonable accuracy are consistently achieved for relevant well quantities. Runtime speedups of around a factor of 370 relative to full- order AD-GPRS simulations are achieved, though the preprocessing needed for POD-TPWL model construction corresponds to the computational requirements for about 2.3 full-order simulation runs. A preliminary treatment for POD-TPWL modeling in which test cases differ from training runs in terms of geological parameters (rather than well controls) is also presented. Results in this case involve only small differences between training and test runs, though they do demonstrate that the approach is able to capture basic solution trends. The impact of some of the detailed numerical treatments within the POD-TPWL formulation is considered in an Appendix.

Mishra, Srikanta↗

DOE OSTI2020

Multiphysics for nuclear energy applications using a cohesive computational framework

With the recent development of advanced numerical algorithms, software design, and low-cost high-performance computer hardware, reliance on coupled multiphysics to predict the behavior of complex physical systems is beginning to become standard practice. This is especially true in nuclear energy applications where strong nonlinear interdependencies exist between reactor physics, radiation transport, multi-scale nuclear fuels performance, thermal fluids, etc. Resolving these nonlinear dependencies requires choices in multiphysics software approaches. Two main multiphysics modeling and simulation approaches have emerged. The first is based upon "code coupling" where disparate physics codes of different software design, code languages, and spatial and temporal integration schemes are coupled together with relatively complex data passing interfaces. The second multiphysics software approach is to employ a "cohesive" framework where all physics applications are developed with a common software design, i.e., data structures, syntax, input format, integrated spatial and temporal discretization schemes, etc. In this paper we present the Multiphysics Object-Oriented Simulation Environment (MOOSE) development and runtime framework and describe the framework's cohesive modeling and simulation multiphysics approach. Then, a "cohesive-like" extension of the MOOSE framework is presented where MOOSE-based physics software applications are efficiently coupled to non-MOOSE (external) physics codes to form multiphysics applications using MOOSE's unique interface capabilities. Finally, several examples of MOOSE's cohesive and cohesive-like multiphysics applications will be demonstrated. These multiphysics demonstrations will incorporate both MOOSE-based applications and external codes, including Nek5000, RELAP-7, TRACE, BISON, and Pronghorn.

22 GENERAL STUDIES OF NUCLEAR REACTORS↗

DOE OSTI2017

Cosmological neutrino simulations at extreme scale

Constraining neutrino mass remains an elusive challenge in modern physics. Precision measurements are expected from several upcoming cosmological probes of large-scale structure. Achieving this goal relies on an equal level of precision from theoretical predictions of neutrino clustering. Numerical simulations of the non-linear evolution of cold dark matter and neutrinos play a pivotal role in this process. We incorporate neutrinos into the cosmological N-body code CUBEP3M and discuss the challenges associated with pushing to the extreme scales demanded by the neutrino problem. We highlight code optimizations made to exploit modern high performance computing architectures and present a novel method of data compression that reduces the phase-space particle footprint from 24 bytes in single precision to roughly 9 bytes. We scale the neutrino problem to the Tianhe-2 supercomputer and provide details of our production run, named TianNu, which uses 86% of the machine (13,824 compute nodes). With a total of 2.97 trillion particles, TianNu is currently the world’s largest cosmological N-body simulation and improves upon previous neutrino simulations by two orders of magnitude in scale. We finish with a discussion of the unanticipated computational challenges that were encountered during the TianNu runtime.

79 ASTRONOMY AND ASTROPHYSICS↗

DOE OSTI2025

Development of a River Dynamical Core for E3SM to simulate compound flooding on Exascale-class heterogeneous supercomputers

Flooding events pose significant risk to human life, property, and infrastructure. Physically-consistent quantification of altered flood risks in global models requires hyper-resolution (~1 km) or fine flood simulations using two-dimensional (2D) physics schemes, both of which are unavailable in the current generation Earth System Models. Here, in this work, we have developed the River Dynamical Core (RDycore), which is an open-source, 2D shallow water equation (SWE) library for the U.S. Department of Energy's Energy Exascale Earth System Model (E3SM). RDycore uses PETSc and libCEED libraries that allows it to run efficiently on CPUs and GPUs, as well as select a time-integration algorithm at runtime without requiring any code modifications. RDycore achieves spatial error convergence rates for problems with analytical and manufactured solutions similar to those reported previously in the literature, or consistent with the implemented first-order spatial discretization scheme. RDycore's accuracy in predicting flooding for a well-studied dam break problem is comparable to existing SWE models. For a problem with 471 million grid cells, RDycore achieves a speedup of 6.6x and 7.6x on GPUs compared to CPUs when using 320 compute nodes on DOE's Perlmutter and Frontier supercomputers, respectively. The one-way coupling of the RDycore library within E3SM is demonstrated by performing multiple 5-day flooding simulations during Hurricane Harvey driven by five precipitation datasets. The E3SM--RDycore simulations at 30 m spatial resolution accurately simulate maximum water height during the hurricane when benchmarked against a previously published study and achieve a speedup of 15x (Perlmutter) and 21x (Frontier) on GPUs relative to CPUs. The work presented here is the foundational step in providing hardware and algorithmic portability framework for simulating kilometer-scale river dynamics within E3SM.

Flood Simulation↗

DOE OSTI2015

Data-dependence Profiling to Enable Safe Thread Level Speculation

Data-dependence profling is a technique that enables a com- piler to judiciously decide when the execution of a loop | which the compiler could not prove to be dependence free | should be speculated through the use of Thread Level Spec- ulation (TLS). The data collected by a data-dependence pro- fler can be used to predict if may dependencies reported by a compiler static analysis are likely to materialize at runtime. A cost analysis can then be used to decide that some loops with a lower probability of dependence should be specula- tively parallelized. This paper addresses the question as to whether a loops' dependence behaviour changes when the in- put to the program changes | a study of 57 different bench- marks indicates that it usually does not change. Then the paper describes SpecEval, an automatic speculative paral- lelization framework that uses single-input data-dependence profles to find speculation candidates in the SPEC2006 and PolyBench/C benchmarks. This paper also presents a per- formance evaluation of TLS implementation in IBM's Blue- Gene/Q supercomputer and shows that the performance of TLS is affected by several factors, including the number of speculated loops, the execution-time coverage of speculated loops, the miss-speculation overhead, the L1 cache miss rate and the effect on dynamic instruction path length.

Bhattacharyya, Arnamoy↗

DOE OSTI2026

Benchmark for two-dimensional large scale coherent structures in partially magnetized E × B plasmas—community collaboration & lessons learned

Low-temperature plasmas (LTPs) are essential to both fundamental scientific research and critical industrial applications. As in many areas of science, numerical simulations have become a vital tool for uncovering new physical phenomena and guiding technological development. Code benchmarking remains crucial for verifying implementations and evaluating performance. This work continues the Landmark benchmark initiative, a series specifically designed to support the verification of LTP codes. In this study, seventeen simulation codes from a collaborative community of nineteen international institutions modeled a partially magnetized E × B Penning discharge. The emergence of large scale coherent structures, or rotating plasma spokes, endows this configuration with an enormous range of time scales, making it particularly challenging to simulate. The codes showed excellent agreement on the rotation frequency of the spoke as well as key plasma properties, including time-averaged ion density, plasma potential, and electron temperature profiles. Achieving this level of agreement came with challenges, and we share lessons learned on how to conduct future benchmarking campaigns. Comparing code implementations, computational hardware, and simulation runtimes also revealed interesting trends, which are summarized with the aim of guiding future plasma simulation software development.

benchmarking↗

DOE OSTI2023

Informed Feature Selection for Data Clustering of CSP Plant Production

To make concentrating solar power (CSP) more cost competitive, rigourous optimizations must be run to improve plant design and operations. However, these optimizaitons rely on time consuming annual simulations that solve an electricity dispatch scheduling problem to maximize plant revenue. To reduce the runtime of annual dispatch simulations of CSP plants, a data clustering approach is utilized. This approach assumes that like days of revenue and electricity generation can be identified using weather and price data. Although weather and price are important factors for electricity production, this work investigates how thermal energy storage (TES) inventory at the beginning of a day, denoted as Si, can be used as a supplemental feature to group like days. A framework for creating and training a deep neural network to predict Si is proposed. This model is validated and assessed using eleven sets of testing data that were not used during training. Then, the data clustering approach is performed three seperate times with features of weather and price along with either Si from the neural network, Si from the full annual simulation, or no Si. Ultimately, the results suggest that using Si as an additional clustering feature improves the data clustering simulation accuracy by 1.4%.

Tuman, Matthew J. (ORCID:000900038772051X)↗

DOE OSTI2018

Synchronization-Avoiding Graph Algorithms

Because they were developed for optimal sequential complexity, classical graph algorithms as found in textbooks have strictly-defined orders of operations. Enforcing a prescribed order of operations, or even an approximate order, in a distributed memory setting requires significant amounts of synchronization, which in turn can severely limit scalability. As a result, new algorithms are typically required to achieve scalable performance, even for solving well-known graph problems. Yet, even in these cases, parallel graph algorithms are written according to parallel programming models that evolved for, e.g., scientific computing, and that still have inherent, and scalability-limiting, amounts of synchronization. In this paper we present a new approach to parallel graph algorithms: synchronization-avoiding algorithms. To eliminate synchronization and its associated overhead, synchronization-avoiding algorithms perform work in an unordered and fully asynchronous fashion in such a way that the result is constantly refined toward its final state. “Wasted” work is minimized by locally prioritizing tasks using problem-dependent task utility metrics. We classify algorithms for graph applications into two broad categories: algorithms with monotonic updates (which evince global synchronization) and algorithms with nonmonotonic updates (which evince vertex-centric synchronization). We apply our approach to both classes and develop novel, synchronization-avoiding algorithms for solving exemplar problems: SSSP and connected components for the former, graph coloring for the latter. We demonstrate that eliminating synchronization in conjunction with effective scheduling policies and optimizations in the runtime results in improved scalability for both classes of algorithms.

Firoz, Jesun S.↗

DOE OSTI2020

Implementing Software Resiliency in HPX for Extreme Scale Computing

The DOE Office of Science Exascale Computing Project (ECP) outlines the next milestones in the supercomputing domain. The target computing systems under the project will deliver 10x performance while keeping the power budget under 30 megawatts. With such large machines, the need to make applications resilient has become paramount. The benefits of adding resiliency to mission critical and scientific applications, includes the reduced cost of restarting the failed simulation both in terms of time and power. Most of the current implementation of resiliency at the software level makes use of a Coordinated Checkpoint and Restart (C/R). This technique of resiliency generates a consistent global snapshot, also called a checkpoint. Generating snapshots involves global communication and coordination and is achieved by synchronizing all running processes. The generated checkpoint is then stored in some form of persistent storage. On failure detection, the runtime initiates a global rollback to the most recent previously saved checkpoint. This involves aborting all running processes, rolling them back to the previous state and restarting them.

97 MATHEMATICS AND COMPUTING↗

DOE OSTI2021

High-Level Synthesis of Parallel Specifications Coupling Static and Dynamic Controllers

The increased need for efficient ways to implement domain-specific accelerators is driving design methodologies towards the use of abstractions higher than the Register Transfer Level (RTL). In this scenario, High Level Synthesis (HLS) plays a significant role by enabling the automatic generation of custom hardware accelerators starting from high level descriptions (e.g., C code). Conventional HLS tools exploit parallelism mostly at the Instruction Level (ILP). They statically schedule the input specifications, and build centralized Finite State Machine (FSM) controllers. However, aggressive exploitation of ILP in many applications has diminishing returns and, usually, centralized approaches do not efficiently exploit coarser parallelism because FSMs are inherently serial. In this paper we present a HLS framework able to synthesize applications that, beside ILP, also expose Task Level Parallelism (TLP). An application can expose TLP through annotations that identify the parallel functions (i.e., tasks). To generate accelerators that efficiently execute concur- rent tasks, we need to solve several issues: devise a mechanism to support concurrent execution flows, exploit memory parallelism, and manage synchronization. To support concurrent execution flows, we introduce a novel adaptive controller. The adaptive controller is composed of a set of interacting control elements that independently manage the execution of a single operation or function call. These control elements check dependencies and resource constraints at runtime, enabling as soon as possible execution. To support parallel access to shared memories and synchronization, we introduce a novel Hierarchical Memory Interface (HMI). With respect to previous solutions, the proposed interface supports multi-ported memories and atomic memory operations, which commonly occur in parallel programming. Our framework can generate the hardware implementation of C functions by employing two different approaches, depending on its characteristics. If a function exposes TLP, then the framework generates hardware implementations based on the adaptive controller. Otherwise, the framework implements the function by exploiting a more conventional FSM approach, which is optimized for ILP exploitation. We evaluate our framework on a set of parallel applications, and show substantial performance improvements (average speedup of 4.7) with limited area over- heads (average area increase of 5.48 times).

Castellana, Vito G.↗