In recent years computer applications have increased in their computational complexity. The industry-wide usage of performance benchmarks, such as SPECmarks, and the popularity of 3 0 graphics applications forces processor designers to pay particular attention to implementation of the floating point unit, or FPU. This paper presents results of the Stanford subnanosecond arithmetic processor (SNAP) research efsort in the design of hardware for floating point addition, multiplication and division. We show that one cycle FP addition is achievable 32% of the time using a variable latency algorithm. For multiplication, a binary tree is often inferior to a Wallace-tree designed using an algorithmic layout approach for contemporary feature sizes (0.3pm). Furthel; in most cases two-bit Booth encoding of the multiplier is preferable to non-Booth encoding for partial product generation. It appears that for division, optimum area-performance is achieved using functional iteration, and we present two techniques to further reduce average division latency.
In this paper intrusion detection using Bayesian probability is discussed. The systems designed are trained a priori using a subset of the KDD dataset. The trained classifier is then tested using a larger subset of KDD dataset. Initially, a system was developed using a naive Bayesian classifier that is used to identify possible intrusions. This classifier was able to detect intrusion with an acceptable detection rate. The classier was then extended to a multi-layer Bayesian based intrusion detection. Finally, we introduce the concept that the best possible intrusion detection system is a layered approach using different techniques in each layer.
A simple flooding mechanism for broadcasting messages over vehicular ad-hoc networks causes massive message redundancy, contention and collision known as the broadcast storm problem. Most available solutions to broadcast storm are normally based on the selection of the rebroadcast nodes depending on a given single attribute such as: sender-to-receiver distance, node density, vehicle's speed, movement direction, or number of message duplicates received. As the settings of any given attributes may remarkably impact the network performance, comparisons of such settings are therefore, an important research topic. In this paper, we suggest a simple but effective solution to the VANET broadcast storm problem through comparison of various attribute settings and select the rebroadcasting nodes based on a combination of multi-network and node attributes. To the best of our knowledge, most of the solutions to the broadcast-storm problem make use of only single attribute schemes and do not consider on employing two or more attributes for broadcast suppression. The scheme suggested throughout this paper is valuable for two reasons: First, it is considered as a guide to select proper methods/attributes for a specific network application's goal, e.g. maximum reachability with a reduced redundancy and/or with minimum message delay at specific network densities. Second, it sets a basis for future studies that employ multiple attributes on mitigating broadcast storm through machine-learning-based approach. The simulation results of this proposed scheme show that a proper combination of attributes may outperform the performance of the best single attribute-based schemes. A combined broadcast scheme which implements both distance and duplicate attributes provides a remarkable reduction on message redundancy by 41 % comparedto the best single distance-based attribute broadcast.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.