arXiv subjects

Saurabh Jha

Publications and source records attributed to Saurabh Jha.

At least 55 records · Page 3Linked to original sources

arXiv2021

Watch out for the risky actors: Assessing risk in dynamic environments for safe driving

Driving in a dynamic environment that consists of other actors is inherently a risky task as each actor influences the driving decision and may significantly limit the number of choices in terms of navigation and safety plan. The risk encountered by the Ego actor depends on the driving scenario and the uncertainty associated with predicting the future trajectories of the other actors in the driving scenario. However, not all objects pose a similar risk. Depending on the object's type, trajectory, position, and the associated uncertainty with these quantities; some objects pose a much higher risk than others. The higher the risk associated with an actor, the more attention must be directed towards that actor in terms of resources and safety planning. In this paper, we propose a novel risk metric to calculate the importance of each actor in the world and demonstrate its usefulness through a case study.

cs.AI↗

arXiv2021

BayesPerf: Minimizing Performance Monitoring Errors Using Bayesian Statistics

Hardware performance counters (HPCs) that measure low-level architectural and microarchitectural events provide dynamic contextual information about the state of the system. However, HPC measurements are error-prone due to non determinism (e.g., undercounting due to event multiplexing, or OS interrupt-handling behaviors). In this paper, we present BayesPerf, a system for quantifying uncertainty in HPC measurements by using a domain-driven Bayesian model that captures microarchitectural relationships between HPCs to jointly infer their values as probability distributions. We provide the design and implementation of an accelerator that allows for low-latency and low-power inference of the BayesPerf model for x86 and ppc64 CPUs. BayesPerf reduces the average error in HPC measurements from 40.1% to 7.6% when events are being multiplexed. The value of BayesPerf in real-time decision-making is illustrated with a simple example of scheduling of PCIe transfers.

cs.DC↗

arXiv2021

Application-aware Congestion Mitigation for High-Performance Computing Systems

High-performance computing (HPC) systems frequently experience congestion leading to significant application performance variation. However, the impact of congestion on application runtime differs from application to application depending on their network characteristics (such as bandwidth and latency requirements). We leverage this insight to develop Netscope, an automated ML-driven framework that considers those network characteristics to dynamically mitigate congestion. We evaluate Netscope on four Cray Aries systems, including a production supercomputer on real scientific applications. Netscope has a lower training cost and accurately estimates the impact of congestion on application runtime with a correlation between 0.7and 0.9 for common scientific applications. Moreover, we find that Netscope reduces tail runtime variability by up to 14.9 times while improving median system utility by 12%.

cs.DC↗

arXiv2020

PS15cey and PS17cke: prospective candidates from the Pan-STARRS Search for Kilonovae

Time domain astronomy was revolutionised with the discovery of the first kilonova, AT2017gfo, in August 2017 which was associated with the gravitational wave signal GW170817. Since this event, numerous wide-field surveys have been optimising search strategies to maximise their efficiency of detecting these fast and faint transients. With the Panoramic Survey Telescope and Rapid Response System (Pan-STARRS), we have been conducting a volume limited survey for intrinsically faint and fast fading events to a distance of $D\simeq200$ Mpc. Two promising candidates have been identified from this archival search, with sparse data - PS15cey and PS17cke. Here we present more detailed analysis and discussion of their nature. We observe that PS15cey was a luminous, fast declining transient at 320 Mpc. Models of BH-NS mergers with a very stiff equation of state could possibly reproduce the luminosity and decline but the physical parameters are extreme. A more likely scenario is that this was a SN2018kzr-like merger event. PS17cke was a faint and fast declining event at 15 Mpc. We explore several explosion scenarios of this transient including models of it as a NS-NS and BH-NS merger, the outburst of a massive luminous star, and compare it against other known fast fading transients. Although there is uncertainty in the explosion scenario due to difficulty in measuring the explosion epoch, we find PS17cke to be a plausible kilonova candidate from the model comparisons.

astro-ph.HE↗

arXiv2020

FIRM: An Intelligent Fine-Grained Resource Management Framework for SLO-Oriented Microservices

Modern user-facing latency-sensitive web services include numerous distributed, intercommunicating microservices that promise to simplify software development and operation. However, multiplexing of compute resources across microservices is still challenging in production because contention for shared resources can cause latency spikes that violate the service-level objectives (SLOs) of user requests. This paper presents FIRM, an intelligent fine-grained resource management framework for predictable sharing of resources across microservices to drive up overall utilization. FIRM leverages online telemetry data and machine-learning methods to adaptively (a) detect/localize microservices that cause SLO violations, (b) identify low-level resources in contention, and (c) take actions to mitigate SLO violations via dynamic reprovisioning. Experiments across four microservice benchmarks demonstrate that FIRM reduces SLO violations by up to 16x while reducing the overall requested CPU limit by up to 62%. Moreover, FIRM improves performance predictability by reducing tail latencies by up to 11x.

cs.DC↗

arXiv2020

Inductive-bias-driven Reinforcement Learning For Efficient Schedules in Heterogeneous Clusters

The problem of scheduling of workloads onto heterogeneous processors (e.g., CPUs, GPUs, FPGAs) is of fundamental importance in modern data centers. Current system schedulers rely on application/system-specific heuristics that have to be built on a case-by-case basis. Recent work has demonstrated ML techniques for automating the heuristic search by using black-box approaches which require significant training data and time, which make them challenging to use in practice. This paper presents Symphony, a scheduling framework that addresses the challenge in two ways: (i) a domain-driven Bayesian reinforcement learning (RL) model for scheduling, which inherently models the resource dependencies identified from the system architecture; and (ii) a sampling-based technique to compute the gradients of a Bayesian model without performing full probabilistic inference. Together, these techniques reduce both the amount of training data and the time required to produce scheduling policies that significantly outperform black-box approaches by up to 2.2x.

cs.DC↗

arXiv2020

ML-driven Malware that Targets AV Safety

Ensuring the safety of autonomous vehicles (AVs) is critical for their mass deployment and public adoption. However, security attacks that violate safety constraints and cause accidents are a significant deterrent to achieving public trust in AVs, and that hinders a vendor's ability to deploy AVs. Creating a security hazard that results in a severe safety compromise (for example, an accident) is compelling from an attacker's perspective. In this paper, we introduce an attack model, a method to deploy the attack in the form of smart malware, and an experimental evaluation of its impact on production-grade autonomous driving software. We find that determining the time interval during which to launch the attack is{ critically} important for causing safety hazards (such as collisions) with a high degree of success. For example, the smart malware caused 33X more forced emergency braking than random attacks did, and accidents in 52.6% of the driving simulations.

cs.CR↗

arXiv2019

Live Forensics for Distributed Storage Systems

We present Kaleidoscope an innovative system that supports live forensics for application performance problems caused by either individual component failures or resource contention issues in large-scale distributed storage systems. The design of Kaleidoscope is driven by our study of I/O failures observed in a peta-scale storage system anonymized as PetaStore. Kaleidoscope is built on three key features: 1) using temporal and spatial differential observability for end-to-end performance monitoring of I/O requests, 2) modeling the health of storage components as a stochastic process using domain-guided functions that accounts for path redundancy and uncertainty in measurements, and, 3) observing differences in reliability and performance metrics between similar types of healthy and unhealthy components to attribute the most likely root causes. We deployed Kaleidoscope on PetaStore and our evaluation shows that Kaleidoscope can run live forensics at 5-minute intervals and pinpoint the root causes of 95.8% of real-world performance issues, with negligible monitoring overhead.

cs.DC↗

arXiv2019

A Study of Network Congestion in Two Supercomputing High-Speed Interconnects

Network congestion in high-speed interconnects is a major source of application run time performance variation. Recent years have witnessed a surge of interest from both academia and industry in the development of novel approaches for congestion control at the network level and in application placement, mapping, and scheduling at the system-level. However, these studies are based on proxy applications and benchmarks that are not representative of field-congestion characteristics of high-speed interconnects. To address this gap, we present (a) an end-to-end framework for monitoring and analysis to support long-term field-congestion characterization studies, and (b) an empirical study of network congestion in petascale systems across two different interconnect technologies: (i) Cray Gemini, which uses a 3-D torus topology, and (ii) Cray Aries, which uses the DragonFly topology.

cs.DC↗

arXiv2019

Understanding Fault Scenarios and Impacts through Fault Injection Experiments in Cielo

We present a set of fault injection experiments performed on the ACES (LANL/SNL) Cray XE supercomputer Cielo. We use this experimental campaign to improve the understanding of failure causes and propagation that we observed in the field failure data analysis of NCSA's Blue Waters. We use the data collected from the logs and from network performance counter data 1) to characterize the fault-error-failure sequence and recovery mechanisms in the Gemini network and in the Cray compute nodes, 2) to understand the impact of failures on the system and the user applications at different scale, and 3) to identify and recreate fault scenarios that induce unrecoverable failures, in order to create new tests for system and application design. The faults were injected through special input commands to bring down network links, directional connections, nodes, and blades. We present extensions that will be needed to apply our methodologies of injection and analysis to the Cray XC (Aries) systems.

cs.DC↗

arXiv2019

Kayotee: A Fault Injection-based System to Assess the Safety and Reliability of Autonomous Vehicles to Faults and Errors

Fully autonomous vehicles (AVs), i.e., AVs with autonomy level 5, are expected to dominate road transportation in the near-future and contribute trillions of dollars to the global economy. The general public, government organizations, and manufacturers all have significant concern regarding resiliency and safety standards of the autonomous driving system (ADS) of AVs . In this work, we proposed and developed (a) `Kayotee' - a fault injection-based tool to systematically inject faults into software and hardware components of the ADS to assess the safety and reliability of AVs to faults and errors, and (b) an ontology model to characterize errors and safety violations impacting reliability and safety of AVs. Kayotee is capable of characterizing fault propagation and resiliency at different levels - (a) hardware, (b) software, (c) vehicle dynamics, and (d) traffic resilience. We used Kayotee to study a proprietary ADS technology built by Nvidia corporation and are currently applying Kayotee to other open-source ADS systems.

cs.SE↗

arXiv2019

ML-based Fault Injection for Autonomous Vehicles: A Case for Bayesian Fault Injection

The safety and resilience of fully autonomous vehicles (AVs) are of significant concern, as exemplified by several headline-making accidents. While AV development today involves verification, validation, and testing, end-to-end assessment of AV systems under accidental faults in realistic driving scenarios has been largely unexplored. This paper presents DriveFI, a machine learning-based fault injection engine, which can mine situations and faults that maximally impact AV safety, as demonstrated on two industry-grade AV technology stacks (from NVIDIA and Baidu). For example, DriveFI found 561 safety-critical faults in less than 4 hours. In comparison, random injection experiments executed over several weeks could not find any safety-critical faults

cs.LG↗

arXiv2019

AVFI: Fault Injection for Autonomous Vehicles

Autonomous vehicle (AV) technology is rapidly becoming a reality on U.S. roads, offering the promise of improvements in traffic management, safety, and the comfort and efficiency of vehicular travel. With this increasing popularity and ubiquitous deployment, resilience has become a critical requirement for public acceptance and adoption. Recent studies into the resilience of AVs have shown that though the AV systems are improving over time, they have not reached human levels of automation. Prior work in this area has studied the safety and resilience of individual components of the AV system (e.g., testing of neural networks powering the perception function). However, methods for holistic end-to-end resilience assessment of AV systems are still non-existent.

eess.SP↗

arXiv2019

Detection of Circumstellar Helium in Type Iax Progenitor Systems

We present direct spectroscopic modeling of 44 Type Iax supernovae (SNe Iax) using spectral synthesis code SYNAPPS. We confirm detections of helium emission in the early-time spectra of two SNe Iax: SNe 2004cs and 2007J. These He I features are better fit by a pure-emission Gaussian than by a P-Cygni profile, indicating that the helium emission originates from the circumstellar environment rather than the SN ejecta. Based on the modeling of the remaining 42 SNe Iax, we find no obvious helium features in other SN Iax spectra. However, $\approx 76\%$ of our sample lack sufficiently deep luminosity limits to detect helium emission with a luminosity of that seen in SNe 2004cs and 2007J. Using the objects with constraining luminosity limits, we calculate that 33% of SNe Iax have detectable helium in their spectra. We examine 11 SNe Iax with late-time spectra and find no hydrogen or helium emission from swept up material. For late-time spectra, we calculate typical upper limits of stripped hydrogen and helium to be $2 \times 10^{-3}$ M$_{\odot}$ and $10^{-2}$ M$_{\odot}$, respectively. While detections of helium in SNe Iax support a white dwarf-He star binary progenitor system (i.e., a single-degenerate [SD] channel), non-detections may be explained by variations in the explosion and ejecta material. The lack of helium in the majority of our sample demonstrates the complexity of SN Iax progenitor systems and the need for further modeling. With strong independent evidence indicating that SNe Iax arise from a SD channel, we caution the common interpretation that the lack of helium or hydrogen emission at late-time in SN Ia spectra rules out SD progenitor scenarios for this class.

astro-ph.HE↗

arXiv2019

The Next Generation of Cosmological Measurements with Type Ia Supernovae

While Type Ia Supernovae (SNe Ia) are one of the most mature cosmological probes, the next era promises to be extremely exciting in the number of different ways SNe Ia are used to measure various cosmological parameters. Here we review the experiments in the 2020s that will yield orders of magnitudes more SNe Ia, and the new understandings and capabilities to constrain systematic uncertainties at a level to match these statistics. We then discuss five different cosmological probes with SNe Ia: the conventional Hubble diagram for measuring dark energy properties, the distance ladder for measuring the Hubble constant, peculiar velocities and weak lensing for measuring sigma8 and strong-lens measurements of H0 and other cosmological parameters. For each of these probes, we discuss the experiments that will provide the best measurements and also the SN Ia-related systematics that affect each one.

astro-ph.CO↗

arXiv2019

The Wide Field Infrared Survey Telescope: 100 Hubbles for the 2020s

The Wide Field Infrared Survey Telescope (WFIRST) is a 2.4m space telescope with a 0.281 deg^2 field of view for near-IR imaging and slitless spectroscopy and a coronagraph designed for > 10^8 starlight suppresion. As background information for Astro2020 white papers, this article summarizes the current design and anticipated performance of WFIRST. While WFIRST does not have the UV imaging/spectroscopic capabilities of the Hubble Space Telescope, for wide field near-IR surveys WFIRST is hundreds of times more efficient. Some of the most ambitious multi-cycle HST Treasury programs could be executed as routine General Observer (GO) programs on WFIRST. The large area and time-domain surveys planned for the cosmology and exoplanet microlensing programs will produce extraordinarily rich data sets that enable an enormous range of Archival Research (AR) investigations. Requirements for the coronagraph are defined based on its status as a technology demonstration, but its expected performance will enable unprecedented observations of nearby giant exoplanets and circumstellar disks. WFIRST is currently in the Preliminary Design and Technology Completion phase (Phase B), on schedule for launch in 2025, with several of its critical components already in production.

astro-ph.IM↗

arXiv2014

Optical Spectra of 73 Stripped-Envelope Core-Collapse Supernovae

We present 645 optical spectra of 73 supernovae (SNe) of Types IIb, Ib, Ic, and broad-lined Ic. All of these types are attributed to the core collapse of massive stars, with varying degrees of intact H and He envelopes before explosion. The SNe in our sample have a mean redshift = 4200 km/s. Most of these spectra were gathered at the Harvard-Smithsonian Center for Astrophysics (CfA) between 2004 and 2009. For 53 SNe, these are the first published spectra. The data coverage range from mere identification (1-3 spectra) for a few SNe to extensive series of observations (10-30 spectra) that trace the spectral evolution for others, with an average of 9 spectra per SN. For 44 SNe of the 73 SNe presented here, we have well-determined dates of maximum light to determine the phase of each spectrum. Our sample constitutes the most extensive spectral library of stripped-envelope SNe to date. We provide very early coverage (as early as 30 days before V-band max) for photospheric spectra, as well as late-time nebular coverage when the innermost regions of the SNe are visible (as late as 2 years after explosion, while for SN1993J, we have data as late as 11.6 years). This data set has homogeneous observations and reductions that allow us to study the spectroscopic diversity of these classes of stripped SNe and to compare these to SNe associated with gamma-ray bursts. We undertake these matters in follow-up papers.

astro-ph.HE↗

arXiv2013

A Parallel Simulated Annealing Approach for the Mirrored Traveling Tournament Problem

The Traveling Tournament Problem (TTP) is a benchmark problem in sports scheduling and has been extensively studied in recent years. The Mirrored Traveling Tournament Problem (mTTP) is variation of the TTP that represents certain types of sports scheduling problems where the main objective is to minimize the total distance traveled by all the participating teams. In this paper we test a parallel simulated annealing approach for solving the mTTP using OpenMP on shared memory systems and we found that this approach is superior especially with respect to the number of solution instances that are probed per second. We also see that there is significant speed up of 1.5x - 2.2x in terms of number of solutions explored per unit time.

cs.DC↗