Engineering Papers⌕ Search

Engineering topics

Liu, Frank

Publications and source records attributed to Liu, Frank.

At least 37 records · Page 2

A survey on processing-in-memory techniques: Advances and challenges

Processing-in-memory (PIM) techniques have gained much attention from computer architecture researchers, and significant research effort has been invested in exploring and developing such techniques. Increasing the research activity dedicated to improving PIM techniques will hopefully help deliver PIM’s promise to solve or significantly reduce memory access bottleneck problems for memory-intensive applications. We also believe it is imperative to track the advances made in PIM research to identify open challenges and enable the research community to make informed decisions and adjust future research directions. In this survey, we analyze recent studies that explored PIM techniques, summarize the advances made, compare recent PIM architectures, and identify target application domains and suitable memory technologies. We also discuss proposals that address unresolved issues of PIM designs (e.g., address translation/mapping of operands, workload analysis to identify application segments that can be accelerated with PIM, OS/runtime support, and coherency issues that must be resolved to incorporate PIM). We believe this work can serve as a useful reference for researchers exploring PIM techniques.

97 MATHEMATICS AND COMPUTING↗

IRIS-BLAS: Towards a Performance Portable and Heterogeneous BLAS Library

This paper presents IRIS-BLAS, a novel heterogeneous and performance portable BLAS library. IRIS-BLAS is built on top of the IRIS runtime and multiple vendor and open-source BLAS libraries. It can transparently use all the architectures/devices available in a heterogeneous system, using the appropriate BLAS library based on the task mapping at run time. Thus, IRIS-BLAS is portable across a broad spectrum of architectures and BLAS libraries, alleviating the worry of application developers about modifying the application source code. Even though the emphasis is on portability, IRIS-BLAS provides competitive or even better performance than other state-of-the-art references. Moreover, IRIS-BLAS offers new features such as efficiently using extremely heterogeneous systems composed of multiple GPUs from different hardware vendors.

Miniskar, Narasinga Rao↗

LaRIS: Targeting Portability and Productivity for LAPACK Codes on Extreme Heterogeneous Systems by Using IRIS

In keeping with the trend of heterogeneity in high-performance computing, hardware manufacturers and vendors are developing new architectures and associated software stacks (e.g., libraries) to harness the best possible performance from commonly used kernels (e.g., linear algebra kernels). However, kernels tuned for one architecture are not portable to others. Moreover, the coexistence of different architectures in a single node makes orchestration difficult. To address these challenges, we introduce LaRIS, a portable framework for LAPACK functionalities. LaRIS ensures a separation between linear algebra algorithms and vendor-library kernels by using the IRIS run time and IRIS-BLAS library. Such abstraction at the algorithm level makes the implementation completely agnostic to the vendor library and architecture. LaRIS uses the IRIS run time to dynamically select the vendor-library kernel and suitable processor architecture at run time. Through LU factorization, we demonstrate that LaRIS can fully utilize different heterogeneous systems by launching and orchestrating different vendor-library kernels without any change in the source code.

Monil, M. A. H.↗

Evaluating HPC Kernels for Processing in Memory

Memory subsystems contribute significantly to the performance and energy efficiency of high-performance computing (HPC) applications. Traditional memory technologies with conventional organization (e.g., DRAM) are struggling to keep up with the increasing memory requirements of modern applications. Techniques such as multilayer cache hierarchy and out-of-order execution are still falling short of mitigating the penalty incurred by memory accesses. Processing-in-memory (PIM), which involves moving memory-intensive kernels to memory for execution instead of bringing the data to the processing unit, is emerging as a promising technique. PIM has recently received traction among computer architecture researchers, and the increasing research activity surrounding this technique indicates its potential to alleviate main memory performance bottlenecks. In this paper, we characterize and identify memory-intensive HPC kernels, perform a first-order evaluation of the PIM technique for selected HPC kernels, quantify performance deviation, and analyze the key factors that affect PIM efficiency.

Asifuzzaman, Kazi↗

Gradient-Based Novelty Detection Boosted by Self-Supervised Binary Classification

Novelty detection aims to automatically identify out-of-distribution (OOD) data, without any prior knowledge of them. It is a critical step in data monitoring, behavior analysis and other applications, helping enable continual learning in the field. Conventional methods of OOD detection perform multi-variate analysis on an ensemble of data or features, and usually resort to the supervision with OOD data to improve the accuracy. In reality, such supervision is impractical as one cannot anticipate the anomalous data. In this paper, we propose a novel, self-supervised approach that does not rely on any pre-defined OOD data: (1) The new method evaluates the Mahalanobis distance of the gradients between the in-distribution and OOD data. (2) It is assisted by a self-supervised binary classifier to guide the label selection to generate the gradients, and maximize the Mahalanobis distance. In the evaluation with multiple datasets, such as CIFAR-10, CIFAR-100, SVHN and TinyImageNet, the proposed approach consistently outperforms state-of-the-art supervised and unsupervised methods in the area under the receiver operating characteristic (AUROC) and area under the precision-recall curve (AUPR) metrics. We further demonstrate that this detector is able to accurately learn one OOD class in continual learning.

Sun, Jingbo↗

On the Stochastic Stability of Deep Markov Models

Deep Markov models (DMM) are generative models which are scalable and expressive generalization of Markov models for representation, learning, and inference problems. DMMs using deep neural networks to parametrize the transition of Markov probability distributions have recently been shown to provide more expressiveness in modeling sequential data and dynamical system responses. However, the fundamental stochastic stability guarantees of such models have not been thoroughly investigated. In this paper, we present a rigorous analytical method to prove the necessary and sufficient conditions of DMM's stochastic stability. This task is achieved by spectral analysis of the efficiently computed Jacobians of probabilistic maps modeled by deep neural networks. We make theoretical connections between the eigenvalues of neural network's weights and the different activation function types used on the stability and overall dynamic behavior of DMMs with Gaussian distributions. We empirically substantiate our theoretical results on stochastic stability and eigenvalue spectra via several numerical experiments. Formal stability guarantees of DMMs can substantially improve their robustness and trustworthiness, necessary for reliable use in safety-critical real-world applications.

Drgona, Jan↗

On the Stochastic Stability of Deep Markov Models

Deep Markov models (DMM) are generative models which are scalable and expressive generalization of Markov models for representation, learning, and inference problems. However, the fundamental stochastic stability guarantees of such models have not been thoroughly investigated. In this paper, we present a novel stability analysis method and provide sufficient conditions of DMM's stochastic stability. The proposed stability analysis is based on the contraction of probabilistic maps modeled by deep neural networks. We make connections between the spectral properties of neural network's weights and different types of used activation function on the stability and overall dynamic behavior of DMMs with Gaussian distributions. Based on the theory, we propose a few practical methods for designing constrained DMMs with guaranteed stability. We empirically substantiate our theoretical results via intuitive numerical experiments using the proposed stability constraints.

Drgona, Jan↗

Evolutionary NAS in Light of Model Stability for Accurate Continual Learning

Continual learning, the capability to learn new knowledge from streaming data without forgetting the previous knowledge, is a critical requirement for dynamic learning systems, especially for emerging edge devices such as self-driving cars and drones. However, continual learning is still facing the catastrophic forgetting problem. Previous work illustrate that model performance on continual learning is not only related to the learning algorithms but also strongly dependent on the inherited model, i.e., the model where continual learning starts. The better stability of the inherited model, the less catastrophic forgetting and thus, the inherited model should be elaborately selected. Inspired by this finding, we develop an evolutionary neural architecture search (ENAS) algorithm that emphasizes the Stability of the inherited model, namely ENAS-S. ENAS-S aims to find optimal architectures for accurate continual learning on edge devices. On CIFAR-10 and CIFAR-100, we present that ENAS-S achieves competitive architectures with lower catastrophic forgetting and smaller model size when learning from a data stream, as compared with handcrafted DNNs.

Du, Xiaocong↗

Automated Detection of Instability-Inducing Channel Geometry Transitions in Saint-Venant Simulation of Large-Scale River Networks

A new sweep-search algorithm (SSA) is developed and tested to identify the channel geometry transitions responsible for numerical convergence failure in a Saint-Venant equation (SVE) simulation of a large-scale open-channel network. Numerical instabilities are known to occur at “sharp” transitions in discrete geometry, but the identification of problem locations has been a matter of modeler’s art and a roadblock to implementing large-scale SVE simulations. The new method implements techniques from graph theory applied to a steady-state 1D shallow-water equation solver to recursively examine the numerical stability of each flowpath through the channel network. The SSA is validated with a short river reach and tested by the simulation of ten complete river systems of the Texas–Gulf Coast region by using the extreme hydrological conditions recorded during hurricane Harvey. The SSA successfully identified the problematic channel sections in all tested river systems. Subsequent modification of the problem sections allowed stable solution by an unsteady SVE numerical solver. The new SSA approach permits automated and consistent identification of problem channel geometry in large open-channel network data sets, which is necessary to effectively apply the fully dynamic Saint-Venant equations to large-scale river networks or for city-wide stormwater networks.

54 ENVIRONMENTAL SCIENCES↗

A Hierarchical Task Scheduler for Heterogeneous Computing

Heterogeneous computing is one of the future directions of HPC. Task scheduling in heterogeneous computing must balance the challenge of optimizing the application performance and the need for an intuitive interface with the programming run-time to maintain programming portability. The challenge is further compounded by the varying data communication time between tasks. This paper proposes RANGER, a hardware-assisted task-scheduling framework. By integrating RISC-V cores with accelerators, the RANGER scheduling framework divides scheduling into global and local levels. At the local level, RANGER further partitions each task into fine-grained subtasks to reduce the overall makespan. At the global level, RANGER maintains the coarse granularity of the task specification, thereby maintaining programming portability. The extensive experimental results demonstrate that RANGER achieves a 12.7× performance improvement on average, while only requires 2.7% of area overhead.

Miniskar, Narasinga Rao↗

A Memory Efficient Lock-Free Circular Queue

Hardware queues are import in many applications, such as data transfer, synchronization of concurrent modules with the need of mutual exclusion constructs. State of the art bounded (of a fixed size) lock free circular queues are implemented either by read/write atomic operations, or barrier conditions, or by separating dequeue and enqueue operations. However, these queues always require an unused element at all the times to safe-guard the front and rear pointers of the queue, so as to avoid data race conditions, which leads to the waste of memory. The waste of memory is especially disadvantageous in applications such as I/O data transfer, and image transfer between processing filters, when large element size is needed, We propose a lock-free solution of the bounded circular queue through read/write atomic operations, but without the need of an extra element in the queue. The proposed solution is implemented and verified in both Verilog and ’C’ languages. We also demonstrate its effectiveness by comparing its area and delay metrics with the implementations of other existing designs of queue.

Miniskar, Narasinga Rao↗

Integrating AI with physics-based hydrological models and observations for insightinto changing climate and anthropogenic impacts

Focal Areas: Advanced computational methods that integrate AI, physics, and observations to provide predictive landscape hydrological modeling over large areas (regional, continental, worldwide) while incorporating increasingly available high-resolution data from drones, lidar and satellite. Science Challenge: Landscape data is available at finer scales than can be used in physics-based hydrological (PBH) models for regional or continental terrestrial water modeling. Thus, we throw away observable detail to achieve computability. We argue that integration of AI with PBH models and observed data can be used to provide upscaling for predictive models that are computable, retain physical conservation properties, and represent the fine-scale features that affect complex flow physics through both natural and urban environments. Developing such next-generation capabilities requires outside-the-box thinking that melds the different approaches of AI modeling, PBH modeling, and observation across multiple scales from local drones to satellites.

54 ENVIRONMENTAL SCIENCES↗

A new form of the Saint-Venant equations for variable topography

The solution stability of river models using the one-dimensional (1D) Saint-Venant equations can be easily undermined when source terms in the discrete equations do not satisfy the Lipschitz smoothness condition for partial differential equations. Although instability issues have been previously noted, they are typically treated as model implementation issues rather than as underlying problems associated with the form of the governing equations. This study proposes a new reference slope form of the Saint-Venant equations to ensure smooth slope source terms and eliminate one source of potential numerical oscillations. It is shown that a simple algebraic transformation of channel geometry provides a smooth reference slope while preserving the correct cross-section flow area and the total Piezometric pressure gradient that drives the flow. The reference slope method ensures the slope source term in the governing equations is Lipschitz continuous while maintaining all the underlying complexity of the real-world geometry. The validity of the mathematical concept is demonstrated with the open-source Simulation Program for River Networks (SPRNT) model in a series of artificial test cases and a simulation of a small urban creek. Validation comparisons are made with analytical solutions and the Hydrologic Engineering Center's River Analysis System (HEC-RAS) model. The new method reduces numerical oscillations and instabilities without requiring ad hoc smoothing algorithms.

54 ENVIRONMENTAL SCIENCES↗

Noise-Based Selection of Robust Inherited Model for Accurate Continual Learning

There is a growing demand for an intelligent system to continually learn knowledge from a data stream. Continual learning requires both the preservation of previous knowledge (i.e., avoiding catastrophic forgetting) and the acquisition of new knowledge. Different from previous works that focus only on model adaptation (e.g., regularization, network expansion, memory rehearsal, etc.), we propose a novel training scheme named acquisitive learning (AL), which emphasizes both the knowledge inheritance and knowledge acquisition. AL starts from an elaborately selected model with pre-trained knowledge (the inherited model) and then adapts it to new data using segmented training. The selection is achieved by injecting random noise to various inherited models for better model robustness, which promises higher accuracy in further knowledge acquisition. The approach is validated by the visualization of the loss landscape and quantitative roughness measurement. The combination of the selective inherited model and knowledge acquisition reduces catastrophic forgetting by 10X on the CIFAR-100 dataset.

Du, Xiaocong↗

Deffe: A Data-efficient Framework for Performance Characterization in Domain-Specific Computing

Predicting workload performance is a crucial task for many architecture and system research. In this paper we present De￿e, a framework to estimate the workload performances under varying architectural configurations. The infrastructure component of Deffe is based on scalable and easy-to-use open-source software components. By casting the performance modeling as transfer learning tasks, the modeling component of Deffe can leverage the learned knowledge on one workload, and “transfer” it to a new workload. Extensive experimental results show that the method can achieve superior testing accuracy with an effective reduction of 32-80x in terms of the amount of required training data.

Liu, Frank↗