Abstract: A method is provided for synchronizing local times, maintained at nodes within a network architecture, with a reference time. A node according to the invention can synchronize its local time with the reference time source either actively or in a passive, or eavesdropping, manner. Which of the two manners is to be used preferably depends on whether the node's error exceeds a threshold, and whether the node receives an unsolicited burst of synchronization messages. The active manner is preferably a handshaking scheme, such as probabilistic clock synchronization, in which synchronization is initiated by a request from a node requiring synchronization, and a handshaking exchange of messages between the node and the reference time source establishes a temporal relationship, from which the node produces a synchronized time and a maximum error.

Abstract: A method and apparatus are disclosed for achieving collective consistency in the detection and reporting of failures in a distributed computing system having multiple processors. Each processor is capable of being called by a parallel application for system status. Initially, each processor sends the other processors its view on the status of the processors. It then waits for similar views from other processors except those regarded as failed in its own view. If the received views are identical to the view of the processor, the processor returns its view to the parallel application. In a preferred embodiment, if the views are not identical to its view, the processor sets its view to the union of the received views and its current view. The steps are then repeated. Alternately, the steps are repeated if the processor does not have information that each of the processors not regarded as failed in its view forms an identical union view.

Abstract: The present invention comprises a distributed data processing system including a plurality of data processing elements for expeditiously performing an encoding or prediction function pursuant to a context-based model in an adaptive, optimal and time-progressive manner. The distributed data processing system, having access to each symbol of an input data string at each clock cycle, adaptively generates context-relevant data sets which provide the best model for coding or prediction based on the input symbols. Each symbol and its best model for encoding or prediction emerge concurrently from the system, resulting in a favorable time complexity of .omicron.(n) for an n-symbol input data string.

Abstract: A system is provided for obtaining enhanced quality grayscale printing using standard printer architecture. An image to be printed is scanned to produce several bit maps for image pixels having different optical intensities. The different bit maps are handled by a standard print engine separately, in the same manner as multicolor bit maps would be handled. The image is then printed using a print head assembly having separate print heads for grayscale printing using different optical intensities. Thus, a printed image having enhanced resolution and quality, because of the optical intensity discrimination, is produced using a printer architecture which is the same as a standard multicolor printer architecture except for the front end, which scans for optical intensity, and the print head assembly, which has different print heads designed to produce different optical intensities.

Abstract: A matrix transpose method for transposing any size matrix on a 2-dimensional mesh multi-node system with circuit-switched-like routing in the iterative and recursive forms. The matrix transpose method involves a two-level decomposition technique of first partitioning each mesh on a diagonal axis into four submeshes and then further partitioning each of the four submeshes on the diagonal axis into four submeshes. The transposition of all off-diagonal submatrices can be performed concurrently and the transposition of all successive on-diagonal submatrices can be performed iteratively or recursively.

Abstract: A method, system, and manufacture are provided, for use in connection with data processing and compression, for quantizing a string of data values, such as image data pixel values. The quantization is achieved by grouping the data values, based on their values, into a predetermined number of categories, each category containing the same total number of values. For each category, a value, preferably a mean value of those in the category, is selected as a quantization value. All of the data values in the category arc then represented by the selected quantization value. For data strings having a dependency (that is, the values of one or more of the data values provide information about values of other of the data values), the dependency is modeled by a method in which a modeling algorithm defines contexts in terms of a tree structure, and the basic method of grouping into categories and selecting a quantization value for each category is performed on a per node (i.e., per context) basis.

Abstract: Parallel ML processing of an analog signal in a RLL-coded channel in which (1) vectors for a current state of the channel and the next state of the channel are computed using Walsh transform vector coefficients of the analog signal; (2) current state vectors and next state vectors and values of vectors precomputed in analog matched filters are used to generate vector scalar products which are compared against preselected threshold values for generating binary decision outputs that are used in digital sequential finite state machines to generate ML symbol decisions; and (3) ML symbol decisions are fed back and used to subtract the intersymbol interference value of the current state vector from the vector of the next state to transform the next state vector into an updated current state vector.

Abstract: A method is provided for providing a cache architecture for a database system having a given number of request streams and a given number of pages of random access memory available for use in one or more caches. The cache architecture includes (i) an allocation of memory pages over a number of caches, and (ii) an assignment of the request streams to the caches. Given that the number of caches is less than the number of streams, the method according to the invention allocates memory pages to the caches and assigns streams to the caches so as to optimize the memory access hit ratio for a given trace of memory requests from the streams. The method includes obtaining characterization information for the request streams (mean burst sizes and cache depth distributions based on the sequence of requests in the trace), and using the characterization information to predict the hit ratios for proposed superpositions of the request streams.

Abstract: The present invention comprises an image processing system to process an array of ultrasonic scanning image data. The image processing system is capable of performing local attenuation estimations by the use of the scanned data and displaying the estimated local attenuations in clearly identifiable images whereby the results of the ultrasonic scan can be accurately and easily interpreted. The estimation process comprises the steps of macro segmentation, micro segmentation, classification, and filtering wherein various techniques of average thresholds, classification and homogenization are used to minimize measurement noises and determine the pixels within a homogeneous geometrical element.

Abstract: A system and method for simulating a mechanistic kinetic process, such as a chemical process including one or more chemical reactions, over a predetermined time period is provided. The simulation proceeds stochastically, by taking discrete time steps through the time period, and performing events (i.e., chemical reactions), based on the relationship between their probabilities of occurrence and the time steps taken. The system and method of the invention include means or method steps for detecting equilibrium conditions, in which a reaction moves forward and backward with a reaction probability which is high, relative to the probabilities of other reactions being simulated. When a reaction in equilibrium is detected, a probability calculation is made, based on the non-equilibrium reactions being simulated. Time steps are made, and events are simulated, based only on the non-equilibrium reactions.

Abstract: A method is provided for defining a minimally sized set of industrial process experiments which are sufficient to identify optimal levels for factors which go into the process. The method defines levels for the factors for the experiments in terms of a symmetric orthogonal array, which represents a valid Galois field for a number of levels for each factor equal to a prime number. The factors are divided into first and second groups. In accordance with Galois field theory, the required number of experiments is a number sufficient to provide a complete set of permutations of levels for the first group of factors. For each experiment, the levels for the second group of factors are determined based on the levels of the first group of factors. In a preferred embodiment, level symbols which are consecutive integers ranging upward from zero are assigned to the levels for each factor.

Abstract: The present invention comprises a computer system having a plurality of processors configured in an architecture having at least two subgraphs wherein at least a first subgraph and a second subgraph having the same topology and corresponding processors being stepwise complimentary for a data redistribution operation. Each processor of the computer system comprises a plurality of data-blocks and an executable program. The executable program performs the data redistribution operation by first exchanging in parallel the first half of the data-blocks of a processor in the first subgraph with the corresponding processor of the second subgraph. The redistributions of data with the corresponding stepwise complimentary processors are then simultaneously performed utilizing the full bandwidth of the data links. A reverse exchange of the first half of the data blocks between the processors of the first and the second subgraphs are then performed at the end.

Abstract: A computer program product, such as a pre-recorded software program on a floppy disk, is disclosed, for use with a processing and database system for off-loading, to disk controller, the extraction of committed data. Responsive to the pre-recorded program instructions, the system first picks a Commit.sub.-- LSN value and insures all the data modified prior to the Commit.sub.-- LSN value is processed following the DBMS policy of reducing some disk I/Os or not for the modified pages cached in the system. If the policy is not to do disk I/Os for such pages, then the system places the identifiers of those pages in an ignore list. Otherwise, the system writes those pages to disk and empties the ignore list. Afterwards, the system forwards the ignore list and the Commit.sub.-- LSN along with information regarding the data to be processed to the controller.

Abstract: The invention relates to searching for a desired data item within a group of data. Representative samples among a group of data to be searched are displayed, and two representative samples in display are selected to specify a new area of data to be searched. Thus, the portion of the data is bracketed and narrowed down, and the desired data item is found quickly and efficiently.

Abstract: A user interface apparatus and method for a computing or processing system employing a graphical user interface is presented. In order to provide a good user feel, and specifically to avoid a feeling of inertia or sluggishness in cursor motion as the user starts or stops movement, the apparatus and method provide a negative inertia transfer function. That is, the cursor movement signal produced by applying the input parameter to a transfer function in accordance with the invention is related to a sum of (i) the input parameter, and (ii) a signal related to the rate of change of the input parameter. This is preferably implemented as an additive high-pass transfer function. Preferred embodiments can employ analog circuitry or a digitally programmed transfer function algorithm.

Abstract: A method for routing multiple message packets to their respective destinations on a parallel system is disclosed which takes into account the value of the communication start-up time and the transmission time For the specific parallel system. The preferred embodiment involves first selecting a base using the parameters of communication start-up and transmission time and then for each datablock, subtracting the node address from the destination address of the datablock using the modulo-n subtractions to Form a relative offset and representing the relative offset in the pre-selected base before sending the datablocks to their destination nodes in phases, each phase involving scanning the individual i-th digits of the relative offset value and packing those datablocks with identical i-th digits of the relative offset together.

Abstract: A method is provided for producing a mesh representation of an arbitrarily shaped object, for applications such as finite element analysis. The method includes the identifying elements of a mesh, such as a tetrahedral mesh, which are suitable, based on predetermined criteria, for merging into one of a predetermined set of target elements. For instance, if the predetermined set of target elements includes pentahedra and hexahedra, and the method operates on a tetrahedral mesh, then tetrahedra of the mesh are identified based on known ways of decomposing the target elements into tetrahedra. The groups of tetrahedra are identified based on whether they share faces in common, and whether they have faces which share a common edge and which are either coplanar or have an angle between the faces which satisfies a predetermined condition. The latter faces are referred to as quadrilateral pairs. A graph representation of the mesh is used, preferably including nodes representing regions of the mesh (i.e.

Abstract: A method and system are provided for subdividing an arbitrarily shaped object into a collection of geometric elements ("cells") having predefined, simple topologies which facilitate further subdivision into tetrahedra, and which are well suited for further applications, such as finite element calculations. Initially, a representation of the object includes one or more regions, at least one of the regions not meeting the definition of "cell", each region having vertices, edges, and faces. An edge is selected according to a priority scheme, and slices which run through the edge and which also either run through other edges or vertices of the representation, are coplanar with other faces of the representation, or have other predetermined attributes, are considered for use in subdividing the representation. A score, which was initially calculated for the representation, is recalculated for hypothetical subdivisions of the representation incorporating each respective one of such slices.

Abstract: A multi-tiered indexing method is disclosed for a partitioned table in a parallel or distributed database system. A Local Index is created and maintained for each partition of the table and a Coarse Global Index is created and maintained. The Coarse Global Index identifies the indexed partition(s) by partition identifiers (PIDs) and associates the individual Index Key Values with their target partitions so that an access request with a highly partition-selective search predicate on the Index Key can be quickly and easily directed to the target partition(s) for processing. An index maintenance locking protocol is also disclosed which handles the insertion and deletion of index entries and assures the consistency between the Local Index entries and the Coarse Global Index entries during concurrent index accesses by different transactions.

Abstract: The present invention comprises a distributed data processing system including a plurality of data processing elements for expeditiously performing an encoding or prediction function pursuant to a context-based model in an adaptive, optimal and time-progressive manner. The distributed data processing system, having access to each symbol of an input data string at each clock cycle, adaptively generates context-relevant data sets which provide the best model for coding or prediction based on the input symbols. Each symbol and its best model for encoding or prediction emerge concurrently from the system, resulting in a favorable time complexity of O(n) for an n-symbol input data string.