https://www.mdu.se/

In this article, we present a machine-learning approach to predict the execution time of concurrent applications by leveraging isolated execution times and performance monitoring counters. Through extensive experiments across various application pairs, we explore the challenges of modeling execution time in a concurrent environment. This study highlights three progressively refined models, transitioning from simple neural networks to complex Bayesian-optimized ensemble techniques. A key finding highlights the significant role of interference in execution time variability as shared resource contention leads to deviations from isolated behavior. These insights enhance the understanding of the role of scheduler and application dynamics in concurrent environments. 

The ability of a system to maintain the availability of its services for the end user is a crucial indicator of performance, both in terms of infrastructure and serviceability. In other words, the service's availability depends on the system's resilience, which is its ability to handle disruptions in the deployed service. Fault management is crucial to increasing system resilience because it aims to control and recover error conditions. However, the efficiency of the fault management implementation depends on the infrastructure design: hardware-assisted fault management allows the quickest recovery action and a better fault isolation. From what we have written above, we can understand why an efficient infrastructure design and practical implementation of fault management can lead to high system resilience. How can we evaluate the efficiency of fault management and infrastructure design relative to system resilience? This paper proposes a new model for measuring system resilience that considers the different fault management aspects contributing to resilience.

Advancement of industrial systems seek improvements to achieve required level of quality of service and efficient performance management. It is essential though to have better understanding of resource utilization behaviour of applications in execution. Even the expert engineers desire to envision dependencies and impact of one computer resource on the other. For such situations it is significant to know statistical relationship between data sets such as a resource with higher cache demand should not be scheduled together with other cache hungry process at the same time and same core. Performance monitoring data coming from hardware and software is huge and grouping of this time series data based on similar behaviour can display distinguishable execution phases. For benefits like these we opt to choose change point analysis method. By using this method study determined the optimal threshold which can identify more or less same segments for other executions of same application and same event. These segments are then validated with the help of test data. Finally the study provided segment-wise, local, compact statistical model with decent accuracy.

In this paper, we present a last-level cache partitioning controller for multi-core systems. Our objective is to control the Quality of Service (QoS) of applications in multi-core systems by monitoring run-time performance and continuously re-sizing cache partition sizes according to the applications' needs. We discuss two different use-cases; one that promotes application fairness and another one that prioritizes applications according to the system engineers' desired execution behavior. We display the performance drawbacks of maintaining a fair schedule for all system tasks and its performance implications for system applications. We, therefore, implement a second control algorithm that enforces cache partition assignments according to user-defined priorities rather than system fairness. Our experiments reveal that it is possible, with non-instrusive (0.3-0.7\% CPU utilization) cache controlling measures, to increase performance according to setpoints and maintain the QoS for specific applications in an over-saturated system.

We present LLM-shark, a tool for automatic hardware resource-boundness detection and cache-partitioning. Our tool has three primary objectives: First, it determines the hardware resource-boundness of a given application. Secondly, it estimates the initial cache partition size to ensure that the application performance is conserved and not affected by other processes competing for cache utilization. Thirdly, it continuously monitors that the application performance is maintained over time and, if necessary, change the cache partition size. We demonstrate LLM-shark's functionality through a series of tests using six different applications, including a set of feature detection algorithms and two synthetic applications. Our tests reveal that it is possible to determine an application's resource-boundness using a Pearson-correlation scheme implemented in LLM-shark. We propose a scheme to size cache partitions based on the correlation coefficient applications depending on their resource boundness.

In this paper, we describe the creation of resource usage forecasts for applications with unknown execution characteristics, by evaluating different regression processes, including autoregressive, multivariate adaptive regression splines, exponential smoothing, etc. We utilize Performance Monitor Units (PMU) and generate hardware resource usage models for the L-2-cache and the L-3-cache using nine different regression processes. The measurement strategy and regression process methodology are general and applicable to any given hardware resource when performance counters are available. We use three benchmark applications: the SIFT feature detection algorithm, a standard matrix multiplication, and a version of Bubblesort. Our evaluation shows that Multi Adaptive Regressive Spline (MARS) models generate the best resource usage forecasts among the considered models, followed by Single Exponential Splines (SES) and Triple Exponential Splines (TES).

In this paper, we discuss a method for application fingerprinting using conventional hardware and software performance counters. Modern applications are complex and often utilizes a broad spectra of the available hardware resources, where multiple performance counters can be of significant interest. The number of performance counters that can be captured simultaneously is, however, small due to hardware limitations in most modern computers. We propose to mitigate the hardware limitations using an intelligent mechanism that pinpoints the most relevant performance counters for an application's performance. In our proposal, we utilize the Pearson correlation coefficient to rank the most relevant PMU events and filter out events of less relevance to an application's execution. Our ultimate goal is to establish a comparable application fingerprint model using performance counters, that we can use to classify applications. The classification procedure can then be used to determine the type of application's fingerprint, such as malicious software.

In this paper, we evaluate different methods for statistical determination of application resource dependency in multi-core systems. We measure the performance counters of an application during run-time and create a system resource usage profile. We then use the resource profile to evaluate the application dependency on the specific resource. We discuss and evaluate two methods to process the data, including moving average filter and partitioning the data into smaller segments in order to interpret data for correlation calculations. Our aim with this study is to evaluate and create a generalizeable methods for automatic determination of resource dependencies. The final outcome of the methods used in this study is the answer to the question: 'To what resources is this application dependent on?'. The recommendation of this tool will be used in conjunction with our last-level cache partitioning controller (LLC-PC), to make decision if an application should receive last-level cache partition slices.

In this paper we present a methodology to be used for quantifying the level of performance isolation for a multi-core system. We have devised a test that can be applied to breaches of isolation in different computing resources that may be shared between different cores. We use this test to determine the level of isolation gained by using the Jailhouse hypervisor compared to a regular Linux system in terms of CPU isolation, cache isolation and memory bus isolation. Our measurements show that the Jailhouse hypervisor provides performance isolation of local computing resources such as CPU. We have also evaluated if any isolation could be gained for shared computing resources such as the system wide cache and the memory bus controller. Our tests show no measurable difference in partitioning between a regular Linux system and a Jailhouse partitioned system for shared resources. Using the Jailhouse hypervisor provides only a small noticeable overhead when executing multiple shared-resource intensive tasks on multiple cores, which implies that running Jailhouse in a memory saturated system will not be harmful. However, contention still exist in the memory bus and in the system-wide cache.