Mark A. Franklin

CommBench-a telecommunications benchmark for network processors

The paper presents a benchmark, CommBench, for use in evaluating and designing telecommunications network processors. The benchmark applications focus on small, computationally intense program kernels typical of the network processor environment. The benchmark is composed of eight programs, four of them oriented towards packet header processing and four oriented towards data stream processing. The benchmark is defined and characteristics such as instruction frequencies, computational complexity, and cache performance are presented. These measured characteristics are compared to the standard SPEC benchmark. Three examples are presented indicating how CommBench can aid in the design of a single chip network multiprocessor.

Optimum buffer circuits for driving long uniform lines

The design of optimum buffer circuits for driving long uniform lines is discussed. Given a uniform line, the size of the buffer driving the line, and the value of the capacitive load driven by the line, the problem considered consists of determining the type, number, and position of buffers that minimize the delay in the line. A variation of this problem that is also considered consists of minimizing the delay in the line when the area occupied by the buffers is constrained; this leads to the solution of the problem of minimizing the delay in driving a pure capacitive load under buffer area constraint. The optimal solution is formally developed, and some very good approximate solutions that can be obtained via simple computations are presented. It is shown that accepting a small increase in delay (of usually 5% over the minimum) can lead to a significant (about 50%) decrease in the area occupied by the buffers. Design curves that allow the reader to determine the optimum buffers with little effort are presented. >

Parallel simulated annealing using speculative computation

A parallel simulated annealing algorithm that is problem-independent, maintains the serial decision sequence, and obtains speedup which can exceed log/sub 2/P on P processors is discussed. The algorithm achieves parallelism by using the concurrency technique of speculative computation. Implementation of the parallel algorithm on a hypercube multiprocessor and application to a task assignment problem are described. The simulated annealing solutions are shown to be, on average, 28% better than the solutions produced by a random task assignment algorithm and 2% better than the solutions produced by a heuristic. >

A workload for evaluating deep packet inspection architectures

High-speed content inspection of network traffic is an important new application area for programmable networking systems, and has recently led to several proposals for high-performance regular expression matching. At the same time, the number and complexity of the patterns present in well-known network intrusion detection systems has been rapidly increasing. This increase is important since both the practicality and the performance of specific pattern matching designs are strictly dependent upon characteristics of the underlying regular expression set. However, a commonly agreed upon workload for the evaluation of deep packet inspection architectures is still missing, leading to frequent unfair comparisons, and to designs lacking in generality or scalability. In this paper, we propose a workload for the evaluation of regular expression matching architectures. The workload includes a regular expression model and a traffic generator, with the former characterizing different levels of expressiveness within rule-sets and the latter characterizing varying degrees of malicious network activity. The proposed workload is used here to evaluate designs (e.g., different memory layouts and hardware organizations) where the matching algorithm is based on compressed deterministic and non deterministic finite automata (DFAs and NFAs).

Biosequence similarity search on the Mercury system

Biosequence similarity search is an important application in modern molecular biology. Search algorithms aim to identify sets of sequences whose extensional similarity suggests a common evolutionary origin or function. The most widely used similarity search tool for biosequences is BLAST, a program designed to compare query sequences to a database. Here, we present the design of BLASTN, the version of BLAST that searches DNA sequences, on the Mercury system, an architecture that supports high-volume, high-throughput data movement off a data store and into reconfigurable hardware. An important component of application deployment on the Mercury system is the functional decomposition of the application onto both the reconfigurable hardware and the traditional processor. Both the Mercury BLASTN application design and its performance analysis are described.

A Learning Identification Algorithm and Its Application to an Environmental System

An empirical heuristic learning identification algorithm of Ivakhnenko was modified and used to model an environmental system producing high nitrate levels in agricultural drain water in the Corn Belt. The method amounts to fitting a polynomial to a multi-input single-output response surface. The modifications result in a reduced number of terms in final model equations, a decrease in computational difficulties, and other improvements in the algorithm. This method appears to be advantageous with systems characterized by complexity with many variables and parameters, ill-defined mathematical structures, and limited data. In other words, this algorithm is useful for empirically generating hypotheses about systems of which relatively little is known.

The Mercury system: exploiting truly fast hardware for data search

In many data mining applications, the size of the database is not only extremely large, it is also growing rapidly. Even for relatively simple searches, the time required to move the data off magnetic media, cross the system bus into main memory, copy into processor cache, and then execute code to perform a search is prohibitive. We are building a system in which a significant portion of the data mining task (i.e., the portion that examines the bulk of the raw data) is implemented in fast hardware, close to the magnetic media on which it is stored. Furthermore, this hardware can be replicated allowing mining tasks to be performed in parallel, thus providing further speedup for the overall mining application. In this paper, we describe a general framework under which this can be accomplished and provide initial performance results for a set of applications.

Auto-pipe and the X language: a pipeline design tool and description language

Auto-Pipe is a tool that aids in the design, evaluation and implementation of applications that can be executed on computational pipelines (and other topologies) using a set of heterogeneous devices including multiple processors and FPGAs. It has been developed to meet the needs arising in the domains of communications, computation on large datasets, and real time streaming data applications. This paper introduces the Auto-Pipe design flow and the X design language, and presents sample applications. The applications include the Triple-DES encryption standard, a subset of the signal-processing pipeline for VERITAS, a high-energy gamma-ray astrophysics experiment. These applications are discussed and their description in X is presented. From X, simulations of alternative system designs and stage-to-device assignments are obtained and analyzed. The complete system permits production of executable code and bit maps that may be downloaded onto real devices. Future work required to complete the Auto-Pipe design tool is discussed.

Design Tradeoffs for Embedded Network Processors

Demands for flexible processing have moved general-purpose processing into the data path of networks. With the development of System-On-a-Chip technology, it is possible to put a number of processors with memory and I/O components on a single ASIC. We present a performance model of such a system and show how the number of processors, cache sizes, and the tradeoffs between the use of on-chip SRAM and DRAM can be optimized in terms of computation per unit chip area for a given workload. Based on a telecommunications benchmark the results of such an optimization are presented and design tradeoffs for Systems-on-a-Chip are identified and discussed.

Statistics on Logic Simulation

The high costs associated with logic simulation of large VLSI based systems have led to the need for new computer architectures tailored to the simulation task. Such architectures have the potential for significant speedups over standard software based logic simulators. Several commercial simulation engines have been produced to satisfy needs in this area. To properly explore the space of alternative simulation architectures, data is required on the simulation process itself. This paper presents a framework for such data gathering activity by first examining possible sources of speedup in the logic simulation task, examining the sort of data needed in the design of simulation engines, and then presenting such data. The data contained in the paper includes information on subtask times found in standard discrete event simulation algorithms, event intensities, queue length distributions and simultaneous event distributions.