Hannes Payer

Distributed queues in shared memory: multicore performance and scalability through quantitative relaxation

A prominent remedy to multicore scalability issues in concurrent data structure implementations is to relax the sequential specification of the data structure. We present distributed queues (DQ), a new family of relaxed concurrent queue implementations. DQs implement relaxed queues with linearizable emptiness check and either configurable or bounded out-of-order behavior or pool behavior. Our experiments show that DQs outperform and outscale in micro- and macrobenchmarks all strict and relaxed queue as well as pool implementations that we considered.

Automated classification of duodenal imagery in celiac disease using evolved Fourier feature vectors

Feature extraction techniques based on selection of highly discriminant Fourier filters have been developed for an automated classification of magnifying endoscope images with respect to pit patterns of colon lesions. These are applied to duodenal imagery for diagnosis of celiac disease. Features are extracted from the Fourier domain by selecting the most discriminant features using an evolutionary algorithm. Subsequent classification is performed with various standard algorithms (KNN, SVM, Bayes classifier) and combination of several Fourier filters and classifiers which is called multiclassifier. The obtained results are promising, due to a high specificity for the detection of mucosal damage typical of untreated celiac disease.

Incorrect systems: it's not the problem, it's the solution

We present an overview of state-of-the-art work in the engineering of digital systems (hardware and software) where traditional correctness requirements are relaxed, usually for higher performance and lower resource consumption but possibly also for other non-functional properties such as more robustness and less cost. The work presented here is categorized into work that involves just hardware, hardware and software, and just software. In particular, we discuss work on probabilistic and approximate design of processors, unreliable cores in asymmetric multi-core architectures, best-effort computing, stochastic processors, accuracy-aware program transformations, and relaxed concurrent data structures. As common theme we identify, at least intuitively, “metrics of correctness” in each piece of work which appear to be important for understanding the effects of relaxed correctness requirements and their relationship to performance improvements and resource consumption.

Memento mori: dynamic allocation-site-based optimizations

Languages that lack static typing are ubiquitous in the world of mobile and web applications. The rapid rise of larger applications like interactive web GUIs, games, and cryptography presents a new range of implementation challenges for modern virtual machines to close the performance gap between typed and untyped languages. While all languages can benefit from efficient automatic memory management, languages like JavaScript present extra thrill with innocent-looking but difficult features like dynamically-sized arrays, deletable properties, and prototypes. Optimizing such languages requires complex dynamic techniques with more radical object layout strategies such as dynamically evolving representations for arrays. This paper presents a general approach for gathering temporal allocation site feedback that tackles both the general problem of object lifetime estimation and improves optimization of these problematic language features. We introduce a new implementation technique where allocation mementos processed by the garbage collector and runtime system efficiently tie objects back to allocation sites in the program and dynamically estimate object lifetime, representation, and size to inform three optimizations: pretenuring, pretransitioning, and presizing. Unlike previous work on pretenuring, our system utilizes allocation mementos to achieve fully dynamic allocation-site-based pretenuring in a production system. We implement all of our techniques in V8, a high performance virtual machine for JavaScript, and demonstrate solid performance improvements across a range of benchmarks.

Performance, scalability, and semantics of concurrent FIFO queues

We introduce the notion of a k-FIFO queue which may dequeue elements out of FIFO order up to a constant k≥0. Retrieving the oldest element from the queue may require up to k+1 dequeue operations (bounded lateness), which may return elements not younger than the k+1 oldest elements in the queue (bounded age) or nothing even if there are elements in the queue. A k-FIFO queue is starvation-free for finite k where k+1 is what we call the worst-case semantical deviation (WCSD) of the queue from a regular FIFO queue. The WCSD bounds the actual semantical deviation (ASD) of a k-FIFO queue from a regular FIFO queue when applied to a given workload. Intuitively, the ASD keeps track of the number of dequeue operations necessary to return oldest elements and the age of dequeued elements. We show that a number of existing concurrent algorithms implement k-FIFO queues whose WCSD are determined by configurable constants independent from any workload. We then introduce so-called Scal queues, which implement k-FIFO queues with generally larger, workload-dependent as well as unbounded WCSD. Since ASD cannot be obtained without prohibitive overhead we have developed a tool that computes lower bounds on ASD from time-stamped runs. Our micro- and macrobenchmarks on a state-of-the-art 40-core multiprocessor machine show that Scal queues, as an immediate consequence of their weaker WCSD, outperform and outscale existing implementations at the expense of moderately increased lower bounds on ASD.

Tempo: Disk drive power consumption characterization and modeling

Accurately modeling a hard disk drive performance and power requirements is of significant interest to designers of Consumer Electronic systems that include a disk drive as a storage device. In this paper, a method to create a detailed power and performance model of a disk is presented. The proposed method is used to accurately estimate the disk power consumption using a trace of all requests issued to the disk with a real-time streaming workload.

Fast and Scalable, Lock-Free k-FIFO Queues

We introduce fast and scalable algorithms that implement bounded- and unbounded-size lock-free k-FIFO queues on parallel, shared memory hardware. Logically, a k-FIFO queue can be understood as queue where elements may be dequeued out-of-order up to k-1, or as pool where the oldest element is dequeued within at most k dequeue operations. The presented algorithms enable up to k enqueue and k dequeue operations to be performed in parallel. Unlike previous designs, however, the algorithms also implement linearizable emptiness and full checks without impairing scalability. We show experimentally that there exist optimal and robust k that result in best access performance and scalability. We then demonstrate that our algorithms outperform and outscale all state-of-the-art concurrent pool and queue algorithms that we considered in all micro- and most macrobenchmarks. Moreover, we demonstrate a prototypical controller which identifies optimal k automatically at runtime achieving better performance than with any statically configuredi¾?k.

Programmable temporal isolation through variable-bandwidth servers

We introduce variable-bandwidth servers (VBS) for scheduling and executing processes under programmable temporal isolation. A VBS is an extension of a constant-bandwidth server where throughput and latency of process execution can not only be controlled to remain constant across different competing workloads but also to vary in time as long as the resulting bandwidth stays below a given bandwidth cap. We have designed and implemented a VBS-based EDF-style constant-time scheduling algorithm, a constant-time admission test, and four alternative queue management plugins which influence the scheduling algorithms overall temporal and spatial complexity. Experiments confirm the theoretical bounds in a number of microbenchmarks and demonstrate that the scheduler can effectively manage in constant time any number of processes up to available memory while maintaining response times of individual processes within a bounded range. We have also developed a small-footprint, bare-metal virtual machine that uses VBS for temporal isolation of multiple, concurrently running processes executing real code.

Scalability versus semantics of concurrent FIFO queues

Maintaining data structure semantics of concurrent queues such as first-in first-out (FIFO) ordering requires expensive synchronization mechanisms which limit scalability. However, deviating from the original semantics of a given data structure may allow for a higher degree of scalability and yet be tolerated by many concurrent applications. We introduce the notion of a k-FIFO queue which may be out of FIFO order up to a constant k (called semantical deviation). Implementations of k-FIFO queues may be distributed and therefore be accessed unsynchronized while still being starvation-free. We show that k-FIFO queues whose implementations are based on state-of-the-art FIFO queues, which typically do not scale under high contention, provide scalability. Moreover, probabilistic versions of k-FIFO queues improve scalability further but only bound semantical deviation with high probability.

Computer-Aided Classification of Zoom-Endoscopical Images Using Fourier Filters

This paper describes an application of machine learning techniques and evolutionary algorithms to colon cancer diagnosis. We propose an automated classification system for endoscopical images, which is supposed to support physicians in making correct decisions. Classification is done according to the pit-pattern scheme, which defines two/six different classes based on the occurrence of patterns on the mucosa. All discriminative information for classification is obtained by filtering an images frequency domain. A major part of this paper is devoted to the search for proper frequency filters. An extensive experimental study compares different search strategies and the resulting classification accuracies. We result in a top classification accuracy of 96.9% and 86.8% for the two- and six-classes case, respectively, using a database of 484 zoom-endoscopic images. We observe a tendency toward the employment of lower frequency filter structures for the best classification settings.