Search NASA⌕ Search

Engineering topics

Chen, Peng

Publications and source records attributed to Chen, Peng.

At least 19 records
All sourcesNASA NTRSDOE OSTI
DOE OSTI2024

autoGEMM: Pushing the Limits of Irregular Matrix Multiplication on Arm Architectures

This paper presents an open-source library that pushes the limits of performance portability for irregular General Matrix Multiplication (GEMM) on the widely-used Arm architectures. Our library, autoGEMM, is designed to support a wide range of Arm processors: from edge devices to HPC-grade CPUs. autoGEMM generates optimized kernels for various hardware configurations by auto-combining fragments of autogenerated micro-kernels that employ hand-written optimizations to maximize computational efficiency. We optimize the kernel pipeline by tuning the register reuse and the data load/store overlapping. In addition, we use a dynamic tiling scheme to generate balanced tile shapes. Finally, we position autoGEMM on top of the TVM framework where our dynamic tiling scheme prunes the search space for TVM to identify the optimal combination of parameters for code optimization. Evaluations on five different classes of Arm chips demonstrate the advantages of autoGEMM. For small matrices, autoGEMM achieves 98% of peak and up to 2.0x speedup over state-of-the-art libraries such as LIBXSMM and LibShalom. For irregular matrices (i.e. tall skinny and long rectangles), autoGEMM is 1.3-2.0x faster than widely-used libraries such as OpenBLAS and Eigen. autoGEMM is publicly available at: https://github.com/wudu98/autoGEMM.

Wu, Du↗
DOE OSTI2024

Adaptive Patching for High-resolution Image Segmentation with Transformers

Attention-based models are proliferating in the space of image analytics, including segmentation. The standard method of feeding images to transformer encoders is to divide the images into patches and then feed the patches to the model as a linear sequence of tokens. For high-resolution images, e.g. microscopic pathology images, the quadratic compute and memory cost prohibits the use of an attention-based model, if we are to use smaller patch sizes that are favorable in segmentation. The solution is to either use custom complex multi-resolution models or approximate attention schemes. We take inspiration from Adapative Mesh Refinement (AMR) methods in HPC by adaptively patching the images, as a pre-processing step, based on the image details to reduce the number of patches being fed to the model, by orders of magnitude. This method has a negligible overhead, and works seamlessly with any attention-based model, i.e. it is a pre-processing step that can be adopted by any attention-based model without friction. We demonstrate superior segmentation quality over SoTA segmentation models for realworld pathology datasets while gaining a geomean speedup of 6.9× for resolutions up to 64K2, on up to 2, 048 GPUs.

Zhang, Enzhi↗
DOE OSTI2024

Real-time High-resolution X-Ray Computed Tomography

Computed Tomography (CT) serves as a key imaging technology that relies on computationally intensive filtering and back-projection algorithms for 3D image reconstruction. While conventional high-resolution image reconstruction (> 2K3) solutions provide quick results, they typically treat reconstruction as an offline workload to be performed remotely on large-scale HPC systems. The growing demand for post-construction AI-driven analytics and the need for real-time adjustments call for high-resolution reconstruction solutions that are feasible on local computing resources, i.e. a multi-GPU server at most. In this paper, we propose a novel approach that utilizes Tensor Cores to optimize image reconstruction without sacrificing precision. We also introduce a framework designed to enable real-time execution of end-to-end distributed image reconstruction in a multi-GPU environment. Evaluations conducted on a single Nvidia A100 and H100 GPU show performance improvements of 1.91 × and 2.15 × compared to highly optimized production libraries. Furthermore, our framework, when deployed on 8-card Nvidia A100 GPU system, demonstrates the ability to reconstruct real-world datasets into 20483 volumes (32 GB) in slightly more than one minute and 40963 volumes (256 GB) in 7 minutes.

Wu, Du↗
DOE OSTI2024

Superionic fluoride gate dielectrics with low diffusion barrier for two-dimensional electronics

Exploration of new dielectrics with a large capacitive coupling is an essential topic in modern electronics when conventional dielectrics suffer from the leakage issue near the breakdown limit. Here, to address this looming challenge, we demonstrate that rare-earth metal fluorides with extremely low ion migration barriers can generally exhibit an excellent capacitive coupling over 20 μF cm -2 (with an equivalent oxide thickness of ~0.15 nm and a large effective dielectric constant near 30) and great compatibility with scalable device manufacturing processes. Such a static dielectric capability of superionic fluorides is exemplified by MoS 2 transistors exhibiting high on/off current ratios over 10 8 , ultralow subthreshold swing of 65 mV dec -1 and ultralow leakage current density of ~10 -6 A cm -2 . Therefore, the fluoride-gated logic inverters can achieve notably higher static voltage gain values (surpassing ~167) compared with a conventional dielectric. Furthermore, the application of fluoride gating enables the demonstration of NAND, NOR, AND and OR logic circuits with low static energy consumption. In particular, the superconductor–insulator transition at the clean-limit Bi 2 Sr 2 CaCu 2 O 8+δ can also be realized through fluoride gating. In conclusion, our findings highlight fluoride dielectrics as a pioneering platform for advanced electronic applications and for tailoring emergent electronic states in condensed matter.

75 CONDENSED MATTER PHYSICS, SUPERCONDUCTIVITY AND↗
DOE OSTI2024

Molecular-scale Insights into Cooperativity Switching of x TAB Adsorption on Gold Nanoparticles

Quantifying adsorption behaviors is crucial for various applications such as catalysis, separation, and sensing, yet it is generally challenging to access in solution. Here, we report a combined experimental and computational study of the adsorption behaviors of alkyl-trimethylammonium bromides (xTAB), a class of ligands important for colloidal nanoparticle stabilization and shape control, with various alkyl chain lengths x on Au nanoparticles. We use density functional theory (DFT) to calculate xTAB binding energies on Au{111} and Au{110} surfaces with standing-up and lying-down configurations, which provides insights into the adsorption affinity and cooperativity differences of xTAB on these two facets. We demonstrate the key role of van der Waals interactions in determining the xTAB adsorption behavior. These computational results predict and explain the experimental discovery of xTAB’s adsorption behavior switch from stronger affinity, negative cooperativity to weaker affinity, positive cooperativity when the concentration of xTAB increases in solution. We also show that in the standing-up configuration, bilayer adsorption may occur on both facets, which can lead to different differential binding energies and consequently adsorption crossover between the two facets when the ligand concentration increases. Our combined experimental and computational approaches demonstrate a paradigm for gaining molecular-scale insights into adsorbate–surface interactions.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗
DOE OSTI2024

Efficient Distributed Sequence Parallelism for Transformer-Based Image Segmentation

We introduce an efficient distributed sequence parallel approach for training transformer-based deep learning image segmentation models. The neural network models are comprised of a combination of a Vision Transformer encoder with a convolutional decoder to provide image segmentation mappings. The utility of the distributed sequence parallel approach is especially useful in cases where the tokenized embedding representation of image data are too large to fit into standard computing hardware memory. To demonstrate the performance and characteristics of our models trained in sequence parallel fashion compared to standard models, we evaluate our approach using a 3D MRI brain tumor segmentation dataset. We show that training with a sequence parallel approach can match standard sequential model training in terms of convergence. Furthermore, we show that our sequence parallel approach has the capability to support training of models that would not be possible on standard computing resources.

Lyngaas, Isaac↗
DOE OSTI2023

Evolutionary Architecture Search for Generative Adversarial Networks Based on Weight Sharing

Generative adversarial networks (GANs) are a powerful generative technique but frequently face challenges with training stability. Network architecture plays a significant role in determining the final output of GANs, but designing a fine architecture demands extensive domain expertise. This article aims to address this issue by searching for high-performance generator’s architectures through neural architecture search (NAS). The proposed approach, called evolutionary weight sharing GANs (EWSGAN), is based on weight sharing and comprises two steps. First, a supernet of the generator is trained using weight sharing. Second, a multiobjective evolutionary algorithm (MOEA) is employed to identify optimal subnets from the supernet. These subnets inherit weights directly from the supernet for fitness assessment. Two strategies are used to stabilize the training of the generator supernet: 1) a fair single-path sampling strategy and 2) a discarding strategy. Experimental results indicate that the architecture searched by our method achieved a new state-of-the-art among NAS–GAN methods with a Fréchet inception distance (FID) of 9.09 and an inception score (IS) of 8.99 on the CIFAR-10 dataset. Finally, it also demonstrates competitive performance on the STL-10 dataset, achieving FID of 21.89 and IS of 10.51.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2023

Interfacial alloying between lead halide perovskite crystals and hybrid glasses

The stellar optoelectronic properties of metal halide perovskites provide enormous promise for next-generation optical devices with excellent conversion efficiencies and lower manufacturing costs. However, there is a long-standing ambiguity as to whether the perovskite surface/interface (e.g. structure, charge transfer or source of off-target recombination) or bulk properties are the more determining factor in device performance. Here we fabricate an array of CsPbI 3 crystal and hybrid glass composites by sintering and globally visualise the property-performance landscape. Our findings reveal that the interface is the primary determinant of the crystal phases, optoelectronic quality, and stability of CsPbI 3 . In particular, the presence of a diffusion “alloying” layer is discovered to be critical for passivating surface traps, and beneficially altering the energy landscape of crystal phases. However, high-temperature sintering results in the promotion of a non-stoichiometric perovskite and excess traps at the interface, despite the short-range structure of halide is retained within the alloying layer. By shedding light on functional hetero-interfaces, our research offers the key factors for engineering high-performance perovskite devices.

36 MATERIALS SCIENCE↗
DOE OSTI2023

Solar, Atmospheric, and Volcanic Impacts on 10 Be Depositions in Greenland and Antarctica During the Last 100 Years

Cosmogenic radionuclides (e.g., 10 Be) from ice cores are a powerful tool for solar reconstructions back in time. However, superimposed on the solar signal, other factors like weather/climate and volcanic influences on 10 Be can complicate the interpretation of 10 Be data. A comprehensive study of 10 Be records over the recent period, when atmospheric 10 Be production and meteorological conditions are relatively well-known, can improve our interpretation of 10 Be records. Here we conduct a systematic study of the production and climate/volcanic signals in Antarctica and Greenland 10 Be records, including a new 10 Be record from the East GReenland Ice-core Project site. Greenland and Antarctica records show significant decreasing trends (5%–6.5%/decade) for 1900–1950, which is comparable with the expected production rate inferred from sunspot observations. By comparing 10 Be records with reanalysis data and atmospheric circulation patterns, 10 Be records from Southern/Southeastern Greenland are significantly correlated with the Scandinavia pattern. Stacking 10 Be records from different locations can enhance the production signal. However, this approach is not always straightforward as uncertainties in some records can lead to a weaker solar signal. A strategy can be employed to select records for the bipolar stack by comparing Greenland records with Antarctica records, assuming the shared signal is a production signal. Finally, we observe significant increases (36%–64%) in 10 Be depositions in Greenland related to the Agung eruption. This large increase in Greenland 10 Be records after the Agung eruption, could be partly explained by the enhanced air mass transport from mid-latitudes coinciding with the decreased precipitation en-route.

10Be↗
DOE OSTI2023

Single-cell multimodal imaging uncovers energy conversion pathways in biohybrids

Microbe–semiconductor biohybrids, which integrate microbial enzymatic synthesis with the light-harvesting capabilities of inorganic semiconductors, have emerged as promising solar-to-chemical conversion systems. Improving the electron transport at the nano–bio interface and inside cells is important for boosting conversion efficiencies, yet the underlying mechanism is challenging to study by bulk measurements owing to the heterogeneities of both constituents. Here we develop a generalizable, quantitative multimodal microscopy platform that combines multi-channel optical imaging and photocurrent mapping to probe such biohybrids down to single- to sub-cell/particle levels. We uncover and differentiate the critical roles of different hydrogenases in the lithoautotrophic bacterium Ralstonia eutropha for bioplastic formation, discover this bacterium’s surprisingly large nanoampere-level electron-uptake capability, and dissect the cross-membrane electron-transport pathways. Furthermore, this imaging platform, and the associated analytical framework, can uncover electron-transport mechanisms in various types of biohybrid, and potentially offers a means to use and engineer R. eutropha for efficient chemical production coupled with photocatalytic materials.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗
DOE OSTI2023

Multimodal single-Cell/ Particle imaging and engineering for energy conversion in bacteria (Final Technical Report)

Hybrid inorganic-microbial systems have emerged as a potentially transformative approach to combine the light-harvesting capability of inorganic semiconductors and the ability of microbes to orchestrate complex chemical transformations. The objective of this collaborative research is to combine quantum materials synthesis, bacterial synthetic biology, and multimodal single-entity imaging to quantitatively study how hybrid QD-bacteria systems convert light to value chemicals at the single-to-sub cell level, with the ultimate goal of gaining insights to guide the engineering of QDs and bacterial genetics for more efficient bioenergy conversion. The final technical report summarizes our achievements toward this objective.

09 BIOMASS FUELS↗
DOE OSTI2023

How to GIWAXS: Grazing Incidence Wide Angle X‐Ray Scattering Applied to Metal Halide Perovskite Thin Films

Abstract The frequency of reports utilizing synchrotron‐based grazing incident wide angle X‐ray scattering (GIWAXS) to study metal halide perovskite thin films has exploded recently, as this technique has proven invaluable for understanding several structure‐property relationships that fundamentally limit optoelectronic performance. The GIWAXS geometry and temporal resolution are also inherently compatible with in situ and operando setups (including ISOS protocols), and a relatively large halide perovskite research community has deployed GIWAXS to unravel important kinetic and dynamic features in these materials. Considering its rising popularity, the aim here is to accelerate the required learning curve for new experimentalists by clearly detailing the underlying analytical concepts which can be leveraged to maximize GIWAXS studies of polycrystalline thin films and devices. Motivated by the vast range of measurement conditions offered, together with the wide variety of compositions and structural motifs available (i.e., from single‐crystal and polycrystalline systems, to quantum dots and layered superlatices), a comprehensive framework for conducting effective GIWAXS experiments is outlined for different purposes. It is anticipated that providing a clear perspective for this topic will help elevate the quality of future GIWAXS studies—which have become routine—and provide the impetus required to develop novel GIWAXS approaches to resolve unsettled scientific questions.

36 MATERIALS SCIENCE↗
DOE OSTI2023

PERKS: a Locality-Optimized Execution Model for Iterative Memory-bound GPU Applications

Iterative memory-bound solvers commonly occur in HPC codes. Typical GPU implementations have a loop on the host side that invokes the GPU kernel as much as time/algorithm steps there are. The termination of each kernel implicitly acts the barrier required after advancing the solution every time step. We propose an execution model for running memory-bound iterative GPU kernels: PERsistent KernelS (PERKS). In this model, the time loop is moved inside persistent kernel, and device-wide barriers are used for synchronization. We then reduce the traffic to device memory by caching subset of the output in each time step in the unused registers and shared memory. PERKS can be generalized to any iterative solver: they largely independent of the solver's implementation. We explain the design principle of PERKS and demonstrate effectiveness of PERKS for a wide range of iterative 2D/3D stencil benchmarks (geomean speedup of 2.12x for 2D stencils and 1.24x for 3D stencils over state-of-art libraries), and a Krylov subspace conjugate gradient solver (geomean speedup of 4.86x in smaller SpMV datasets from SuiteSparse and 1.43x in larger SpMV datasets over a state-of-art library). All PERKS-based implementations available at: https://github.com/neozhang307/PERKS.

Zhang, Lingqi↗
DOE OSTI2023

Revisiting Temporal Blocking Stencil Optimizations

Iterative stencils are used widely across the spectrum of High Performance Computing (HPC) applications. Many efforts have been put into optimizing stencil GPU kernels, given the prevalence of GPU-accelerated supercomputers. To improve the data locality, temporal blocking is an optimization that combines a batch of time steps to process them together. Under the observation that GPUs are evolving to resemble CPUs in some aspects, we revisit temporal blocking optimizations for GPUs. We explore how temporal blocking schemes can be adapted to the new features in the recent Nvidia GPUs, including large scratchpad memory, hardware prefetching, and device-wide synchronization. We propose a novel temporal blocking method, EBISU, which champions low device occupancy to drive aggressive deep temporal blocking on large tiles that are executed tile-by-tile. We compare EBISU with state-of-the-art temporal blocking libraries: STENCILGEN and AN5D. We also compare with state-of-the-art stencil auto-tuning tools that are equipped with temporal blocking optimizations: ARTEMIS and DRSTENCIL. Over a wide range of stencil benchmarks, EBISU achieves speedups up to 2.53x and a geometric mean speedup of 1.49x over the best state-of-the-art performance in each stencil benchmark.

Zhang, Lingqi↗