Cloud GPU servers have become the de facto way for deep learning practitioners to train complex models on large-scale datasets. However, it is challenging to determine the appropriate cluster configuration---e.g., server type and number---for different training workloads while balancing the trade-offs in training time, cost, and model accuracy. Adding to the complexity is the potential to reduce the

Function-as-a-Service (FaaS) is one form of the serverless cloud computing paradigm and is defined through FaaS platforms (e.g., AWS Lambda) executing event-triggered code snippets (i.e., functions). Many studies that empirically evaluate the performance of such FaaS platforms have started to appear but we are currently lacking a comprehensive understanding of the overall domain. In our work, we survey

Benchmarking of CPU resources in WLCG has been based on the HEP-SPEC06 (HS06) suite for over a decade. It has recently become clear that HS06, which is based on real applications from non-HEP domains, no longer describes typical HEP workloads. The aim of the HEP-Benchmarks project is to develop a new benchmark suite for WLCG compute resources, based on real applications from the LHC experiments. By

Mean field models are a popular tool used to analyse load balancing policies. In some exceptional cases the response time distribution of the mean field limit has an explicit form. In most cases it can be computed using either a recursion or a differential equation (for exponential job sizes with mean one). In this paper we study the value of the mean response time $E[R_\lambda]$ as the arrival rate

With the booming of next generation sequencing technology and its implementation in clinical practice and life science research, the need for faster and more efficient data analysis methods becomes pressing in the field of sequencing. Here we report on the evaluation of an optimized germline mutation calling pipeline, HummingBird, by assessing its performance against the widely accepted BWA-GATK pipeline

Motivated by modern parallel computing applications, we consider the problem of scheduling parallel-task jobs with heterogeneous resource requirements in a cluster of machines. Each job consists of a set of tasks that can be processed in parallel, however, the job is considered completed only when all its tasks finish their processing, which we refer to as "synchronization" constraint. Further, assignment

Optimal caching of files in a content distribution network (CDN) is a problem of fundamental and growing commercial interest. Although many different caching algorithms are in use today, the fundamental performance limits of network caching algorithms from an online learning point-of-view remain poorly understood to date. In this paper, we resolve this question in the following two settings: (1) a

Amid unprecedented times caused by COVID-19, healthcare systems all over the world are strained to the limits of, or even beyond, capacity. A similar event is experienced by some healthcare systems regularly, due to for instance seasonal spikes in the number of patients. We model this as a queueing system in heavy traffic (where the arrival rate is approaching the service rate from below) or in overload

One problem with load test quality, almost always overlooked, is the potential for the load generator's user thread pool to sync up and dispatch queries in bunches rather than independently from each other like real users initiate their requests. A spiky launch pattern misrepresents workload flow as well as yields erroneous application response time statistics. This paper describes what a real user

Deep learning models are increasingly used for end-user applications, supporting both novel features such as facial recognition, and traditional features, e.g. web search. To accommodate high inference throughput, it is common to host a single pre-trained Convolutional Neural Network (CNN) in dedicated cloud-based servers with hardware accelerators such as Graphics Processing Units (GPUs). However

We consider scheduling to minimize mean response time of the M/G/k queue with unknown job sizes. In the single-server case, the optimal policy is the Gittins policy, but it is not known whether Gittins or any other policy is optimal in the multiserver case. Exactly analyzing the M/G/k under any scheduling policy is intractable, and Gittins is a particularly complicated policy that is hard to analyze

FFT, FMM, and multigrid methods are widely used fast and highly scalable solvers for elliptic PDEs. However, emerging large-scale computing systems are introducing challenges in comparison to current petascale computers. Recent efforts (Dongarra et al. 2011) have identified several constraints in the design of exascale software that includes massive concurrency, resilience management, exploiting the

The Gatan K2 IS direct electron detector (Gatan Inc., 2018), which was introduced in 2014, marked a watershed moment in the development of cameras for transmission electron microscopy (TEM) (Pan & Czarnik, 2016). Its pixel frequency, i.e. the number of data points (pixels) recorded per second, was two orders of magnitude higher than the fastest cameras available only five years before. Starting from

Assuming hardware is the major constraint for enabling real-time mobile intelligence, the industry has mainly dedicated their efforts to developing specialized hardware accelerators for machine learning and inference. This article challenges the assumption. By drawing on a recent real-time AI optimization framework CoCoPIE, it maintains that with effective compression-compiler co-design, it is possible

Smart contracts have enabled blockchain systems to evolve from simple cryptocurrency platforms, such as Bitcoin, to general transactional systems, such as Ethereum. Catering for emerging business requirements, a new architecture called execute-order-validate has been proposed in Hyperledger Fabric to support parallel transactions and improve the blockchain's throughput. However, this new architecture

We present our recent code modernizations of the of the ab initio molecular dynamics program CPMD (www.cpmd.org) with a special focus on the ultra-soft pseudopotential (USPP) code path. Following the internal instrumentation of CPMD, all time critical routines have been revised to maximize the computational throughput and to minimize the communication overhead for optimal performance. Throughout the

Deploying big-data Machine Learning (ML) services in a cloud environment presents a challenge to the cloud vendor with respect to the cloud container configuration sizing for any given customer use case. OracleLabs has developed an automated framework that uses nested-loop Monte Carlo simulation to autonomously scale any size customer ML use cases across the range of cloud CPU-GPU "Shapes" (configurations

Writing high-performance code requires significant expertise of the programming language, compiler optimizations, and hardware knowledge. This often leads to poor productivity and portability and is inconvenient for a non-programmer domain-specialist such as a Physicist. More desirable is a high-level language where the domain-specialist simply specifies the workload in terms of high-level operations

Edge networks are promising to provide better services to users by provisioning computing and storage resources at the edge of networks. However, due to the uncertainty and diversity of user interests, content popularity, distributed network structure, cache sizes, it is challenging to decide where to place the content, and how long it should be cached. In this paper, we study the utility optimization

UASs form a large part of the fighting ability of the advanced military forces. In particular, these systems that carry confidential information are subject to security attacks. Accordingly, an Intrusion Detection System (IDS) has been proposed in the proposed design to protect against the security problems using the human immune system (HIS). The IDSs are used to detect and respond to attempts to

Many proofs of concept blockchain applications built using Hyperledger Fabric, a permissioned blockchain platform, have recently been transformed into production. However, the performance provided by Hyperledger Fabric is of significant concern for enterprises due to steady growth in network usage. Hence, in this paper, we study the performance achieved in a Fabric network using vertical scaling (i

Consider a setting where Willie generates a Poisson stream of jobs and routes them to a single server that follows the first-in first-out discipline. Suppose there is an adversary Alice, who desires to receive service without being detected. We ask the question: what is the amount of service that she can receive covertly, i.e. without being detected by Willie? In the case where both Willie and Alice

Understanding network and application performance are essential for debugging, improving user experience, and performance comparison. Meanwhile, modern mobile systems are optimized for energy-efficient computation and communications that may limit the performance of network and applications. In recent years, several tools have emerged that analyze network performance of mobile applications in~situ

The past few years have seen a surge of applying Deep Learning (DL) models for a wide array of tasks such as image classification, object detection, machine translation, etc. While DL models provide an opportunity to solve otherwise intractable tasks, their adoption relies on them being optimized to meet latency and resource requirements. Benchmarking is a key step in this process but has been hampered

This paper discusses the requirements for and performance metrics of the the Grid Computing system used to implement the Locus Algorithm to identify optimum pointings for differential photometry of 61,662,376 stars and 23,779 quasars. Initial operational tests indicated a need for a software system to analyse the data and a High Performance Computing system to run that software in a scalable manner

When considering different hardware platforms, not just the time-to-solution can be of importance but also the energy necessary to reach it. This is not only the case with battery powered and mobile devices but also with high-performance parallel cluster systems due to financial and practical limits on power consumption and cooling. Recent developments in hard- and software have given programmers the

This article presents an automatic approach to quickly derive a good solution for hardware resource partition and task granularity for task-based parallel applications on heterogeneous many-core architectures. Our approach employs a performance model to estimate the resulting performance of the target application under a given resource partition and task granularity configuration. The model is used

Devirtualization is a compiler optimization that replaces indirect (virtual) function calls with direct calls. It is particularly effective in object-oriented languages, such as Java or C++, in which virtual methods are typically abundant. We present a novel abstract model to express the lifetimes of C++ dynamic objects and invariance of virtual table pointers in the LLVM intermediate representation

The idea behind the recently introduced "age of information" performance measure of a networked message processing system is that it indicates our knowledge regarding the "freshness" of the most recent piece of information that can be used as a criterion for real-time control. In this foundational paper, we examine two such measures, one that has been extensively studied in the recent literature and

We present nanoBench, a tool for evaluating small microbenchmarks using hardware performance counters on Intel and AMD x86 systems. Most existing tools and libraries are intended to either benchmark entire programs, or program segments in the context of their execution within a larger program. In contrast, nanoBench is specifically designed to evaluate small, isolated pieces of code. Such code is common

Deep learning for computer vision depends on lossy image compression: it reduces the storage required for training and test data and lowers transfer costs in deployment. Mainstream datasets and imaging pipelines all rely on standard JPEG compression. In JPEG, the degree of quantization of frequency coefficients controls the lossiness: an 8 by 8 quantization table (Q-table) decides both the quality

The SYN flood attack is a common attack strategy on the Internet, which tries to overload services with requests leading to a Denial-of-Service (DoS). Highly asymmetric costs for connection setup - putting the main burden on the attackee - make SYN flooding an efficient and popular DoS attack strategy. Abusing the widely used TCP as an attack vector complicates the detection of malicious traffic and

New challenges in Astronomy and Astrophysics (AA) are urging the need for a large number of exceptionally computationally intensive simulations. "Exascale" (and beyond) computational facilities are mandatory to address the size of theoretical problems and data coming from the new generation of observational facilities in AA. Currently, the High Performance Computing (HPC) sector is undergoing a profound

The widescale deployment of Autonomous Vehicles (AV) seems to be imminent despite many safety challenges that are yet to be resolved. It is well known that there are no universally agreed Verification and Validation (VV) methodologies to guarantee absolute safety, which is crucial for the acceptance of this technology. Existing standards focus on deterministic processes where the validation requires

Supporting Ultra-Reliable Low-Latency Communication (URLLC) in the Internet of Things (IoT) era is challenging due to stringent constraints on latency and reliability combined with the simple circuitry of IoT nodes. Diversity is usually required for sustaining the reliability levels of URLLC, but there is an additional delay associated to auxiliary procedures to be considered, specially when communication

Dynamic affinity load balancing of multi-type tasks on multi-skilled servers, when the service rate of each task type on each of the servers is known and can possibly be different from each other, is an open problem for over three decades. The goal is to do task assignment on servers in a real time manner so that the system becomes stable, which means that the queue lengths do not diverge to infinity

Groups of small and medium enterprises (SMEs) can back each other to obtain loans from banks and thus form guarantee networks. If the loan repayment of a small company in the network defaults, its backers are required to repay the loan. Therefore, risk over networked enterprises may cause significant contagious damage. In real-world applications, it is critical to detect top vulnerable nodes in such

This article is primarily meant to present an early case study on using MLIR, a new compiler intermediate representation infrastructure, for high-performance code generation. Aspects of MLIR covered in particular include memrefs, the affine dialect, and polyhedral utilities and pass infrastructure surrounding those. This article is also aimed at showing the role compiler infrastructure could play in

We describe our process for automatic detection of performance changes for a software product in the presence of noise. A large collection of tests run periodically as changes to our software product are committed to our source repository, and we would like to identify the commits responsible for performance regressions. Previously, we relied on manual inspection of time series graphs to identify significant

An estimated 60% of smartphones sold in 2018 were equipped with multiple rear cameras, enabling a wide variety of 3D-enabled applications such as 3D Photos. The success of 3D Photo platforms (Facebook 3D Photo, Holopix, etc) depend on a steady influx of user generated content. These platforms must provide simple image manipulation tools to facilitate content creation, akin to traditional photo platforms

We present PrecisionBatching, a quantized inference algorithm for speeding up neural network execution on traditional hardware platforms at low bitwidths without the need for retraining or recalibration. PrecisionBatching decomposes a neural network into individual bitlayers and accumulates them using fast 1-bit operations while maintaining activations in full precision. PrecisionBatching not only

There are many applications where users seek to explore the impact of the settings of several categorical variables with respect to one dependent numerical variable. For example, a computer systems analyst might want to study how the type of file system or storage device affects system performance. A usual choice is the method of Parallel Sets designed to visualize multivariate categorical variables

Machine learning (ML) needs industry-standard performance benchmarks to support design and competitive evaluation of the many emerging software and hardware solutions for ML. But ML training presents three unique benchmarking challenges absent from other domains: optimizations that improve training throughput can increase the time to solution, training is stochastic and time to solution exhibits high

The geometry automated theorem proving area distinguishes itself by a large number of specific methods and implementations, different approaches (synthetic, algebraic, semi-synthetic) and different goals and applications (from research in the area of artificial intelligence to applications in education). Apart from the usual measures of efficiency (e.g. CPU time), the possibility of visual and/or readable

Deep learning (DL) workloads are moving towards accelerators for faster processing and lower cost. Modern DL accelerators are good at handling the large-scale multiply-accumulate operations that dominate DL workloads; however, it is challenging to make full use of the compute power of an accelerator since the data must be properly staged in a software-managed scratchpad memory. Failing to do so can

Modern machine learning frameworks can train neural networks using multiple nodes in parallel, each computing parameter updates with stochastic gradient descent (SGD) and sharing them asynchronously through a central parameter server. Due to communication overhead and bottlenecks, the total throughput of SGD updates in a cluster scales sublinearly, saturating as the number of nodes increases. In this

The problem of identifying intersections between two sets of d-dimensional axis-parallel rectangles appears frequently in the context of agent-based simulation studies. For this reason, the High Level Architecture (HLA) specification -- a standard framework for interoperability among simulators -- includes a Data Distribution Management (DDM) service whose responsibility is to report all intersections

The difficulty of deploying various deep learning (DL) models on diverse DL hardware has boosted the research and development of DL compilers in the community. Several DL compilers have been proposed from both industry and academia such as Tensorflow XLA and TVM. Similarly, the DL compilers take the DL models described in different DL frameworks as input, and then generate optimized codes for diverse

In this paper, the adoption of an intelligent reflecting surface (IRS) for multiple single-antenna source terminal (ST)-DT pairs in two-hop networks is investigated. Different from the previous studies on IRS that merely focused on tuning the reflection coefficient of all the reflection elements at IRS, in this paper, we consider the true reflection resource management. Specifically, the true reflection

We present the design and optimization of a linear solver on General Purpose GPUs for the efficient and high-throughput evaluation of the marginalized graph kernel between pairs of labeled graphs. The solver implements a preconditioned conjugate gradient (PCG) method to compute the solution to a generalized Laplacian equation associated with the tensor product of two graphs. To cope with the gap between

It is well known that building analytical performance models in practice is difficult because it requires a considerable degree of proficiency in the underlying mathematics. In this paper, we propose a machine-learning approach to derive performance models from data. We focus on queuing networks, and crucially exploit a deterministic approximation of their average dynamics in terms of a compact system

Power system state estimation plays a fundamental and critical role in the energy management system (EMS). To achieve a high performance and accurate system states estimation, a graph computing based distributed state estimation approach is proposed in this paper. Firstly, a power system network is divided into multiple areas. Reference buses are selected with PMUs being installed at these buses for

Age of Information is an interesting metric that captures the freshness of information in the underlying applications. It is a combination of both packets inter-arrival time and packet transmission delay. In recent times, advanced real-time systems rely on this metric for delivering status updates as timely as possible. This paper aims to accomplish optimal transmission scheduling policy to maintain

Millimeter-wave (mmWave) communication is one of the cornerstone innovations of fifth-generation (5G) wireless networks, thanks to the massive bandwidth available in these frequency bands. To correctly assess the performance of such systems, however, it is essential to have reliable channel models, based on a deep understanding of the propagation characteristics of the mmWave signal. In this respect

Emerging applications -- cloud computing, the internet of things, and augmented/virtual reality -- need responsive, available, secure, ubiquitous, and scalable datacenter networks. Network management currently uses simple, per-packet, data-plane heuristics (e.g., ECMP and sketches) under an intelligent, millisecond-latency control plane that runs data-driven performance and security policies. However

We consider the load balancing problem in large-scale heterogeneous systems with multiple dispatchers. We introduce a general framework called Local-Estimation-Driven (LED). Under this framework, each dispatcher keeps local (possibly outdated) estimates of queue lengths for all the servers, and the dispatching decision is made purely based on these local estimates. The local estimates are updated via

LoRaWAN enables massive connectivity for Internet-of-Things applications. Many published works employ stochastic geometry to derive outage models of LoRaWAN over fading channels assuming fixed transmit power and distance-based spreading factor (SF) allocation. However, in practice, LoRaWAN employs the Adaptive Data Rate (ADR) mechanism, which dynamically adjusts SF and transmit power of nodes based

Motivated by medium access control for resource-challenged wireless Internet of Things (IoT), we consider the problem of queue scheduling with reduced queue state information. In particular, we consider a time-slotted scheduling model with $N$ sensor nodes, with pair-wise dependence, such that Nodes $i$ and $i + 1,~0 < i < N$ cannot transmit together. We develop new throughput-optimal scheduling policies

The complexity of modern and upcoming computing architectures poses severe challenges for code developers and application specialists, and forces them to expose the highest possible degree of parallelism, in order to make the best use of the available hardware. The Intel$^{(R)}$ Xeon Phi$^{(TM)}$ of second generation (code-named Knights Landing, henceforth KNL) is the latest many-core system, which

We develop an interface-modeling framework for quality and resource management that captures configurable working points of hardware and software components in terms of functionality, resource usage and provision, and quality indicators such as performance and energy consumption. We base these aspects on partially-ordered sets to capture quality levels, budget sizes, and functional compatibility. This