Abstract: A method and an apparatus for speech controlled navigation of a pointer in a graphical user interface. Previous methods use speech commands like arrow keys of a keyboard and lack user friendly interfaces. The method and apparatus therefore provides a space of discrete position states (quantization) for the pointer which can be navigated only via those discrete positions by means of speech command input. The granularity of the discrete position states can be adapted to the respective application window and the position states can be represented by a system of coordinates where the speech command input is based on absolute or relative coordinates. Advantageously a copy image of the graphical user interface of operation can be provided in front of or beside the actual user interface and a proxy pointer device is displayed on the copy image. In one embodiment, only the copy image comprises the discrete position states, and the speech input commands are only transferred to the copy image.

Abstract: An apparatus and method is disclosed for converting an input signal having frequency related information sustained over a first duration of time into an output signal sustained over a second duration of time at substantially the same first frequency by adding or subtracting to the effective wave length of the output signal. Preferably, the signals are converted in digital form with samples added or subtracted to frequency convert the signal.

Abstract: The present invention provides a system and a method for tracking parameters of a synthesized ?an! audio signal that reduces the amount of processing time without causing any discernible degradation in the sound quality of the audio signal. An audio signal is intelligently divided into multiple time slices and the parameters of the audio signal are tracked over the duration of the time slice. The time slices are selected so that the actual characteristic of the parameters over the duration of the time slice can be easily approximated by performing simple, non-processor intensive steps. The characteristics of various components of an audio signal such as a volume envelope, pitch envelope, low frequency oscillator, MIDI commands controlling the audio signal, and various other inputs are used to identify control points. Adjacent control points are then selected as the start point and end point of a time slice.

Abstract: An information display method by use of a card including an information display section which includes a reversible thermosensitive recording portion capable of recording information reversibly therein with application of heat thereto, comprises the steps of (1) storing first information in the information display section; (2) comparing the first information stored in the information display section with a reference information; (3) displaying a second information in the information display section, based on the comparison of the first information and said reference information, and repeating the steps (1) to (3) by updating the first information and the second information when further information is stored in the information display section.

Abstract: A means and method for use at a date matching event to determine, in a discreet manner, the date preferences of a plurality of participants. The method includes gathering the plurality of participants at the date matching event, and assigning a unique identification code to each participant. The identification code is worn or otherwise displayed in plain sight, being readily observable by a plurality of the other participants. Each applicant may next, if time permits, be provided a temporal period to introduce themselves to the other participants, which is followed by a temporal interval for applicants to interact. After an interval for interaction, the date preferences of each participant are discreetly collected and analyzed to determine all occurring mutual matches. The mutual match results are then delivered to the participants, wherein matched participants may then meet to discuss the arrangements for their date.

Abstract: An integrated system and method is provided for processing a service request for installation, maintenance or repair of a local loop maintained by a telecommunications company and providing locally switched service to a customer premise.

Abstract: A bill pay system wherein participating consumers pay bills to participating billers through a payment network operating according to preset rules. The participating consumers receive bills from participating billers (paper/mail bills, e-mail notices, implied bills for automatic debits) which indicate an amount, and a unique biller identification number. To authorize a remittance, a consumer transmits to its participating bank a bill pay order indicating a payment date, a payment amount, the consumer's account number with the biller, a source of funds and the biller's biller identification number, either directly or by reference to static data containing those data elements. Bank C then submits a payment message to a payment network, and the payment network, which assigns the biller reference numbers, forwards the payment message to the biller's bank.

Abstract: The present invention relates to the use of computerized intelligent agents to facilitate the integration of networked performance of financial transactions with computerized methods of financial accounting. Incorporated into this combined financial transaction/financial accounting system are intelligent agents that automatically analyze the system information to provide users with financial advice. This invention permits the automated performance on-line of a wide variety of financial transactions and integrates these transactions with computerized financial accounting. All of this information is collated and analyzed automatically by intelligent agents, which generate user-specific financial reports, profiles, and advice, and under appropriate conditions take action.

Abstract: The invention provides systems and methods which assist a contractor in producing a competitive bid proposal. In one exemplary embodiment, a method is provided for producing a bid from at least one drawing including various items to be included in the bid. To produce the bid, the method relies upon a computer having a display screen, an entry device and a processor coupled to the display screen and the entry device. According to the method, at least one characteristic of a first and a second item on the drawing are entered into the processor. The processor then selects a fitting which connects the first and the second items based on the entered characteristics. A bid amount is then produced based at least in part on the cost of the first item, the second item and the fitting.

Abstract: A postage metering system is provided with a countdown time. The countdown timer issues a priority interrupt when the timer has timed out which interrupts the control system of the meter in such a manner as to hold the meter in reset preventing operation of the postage meter. The timer is reset to a fixed period upon each funds recharge after the meter has been placed in recharge mode but prior to exposure of the secure accounting memory to recharge information.

Abstract: The invention provides for the use of synaptic strengths to reflect the connectivity of local space to give a topological representation. Synaptic strengths are zeroed and the mobile entity's position is initialized. A probability of firing (P) is calculated from the spatial firing pattern of each cell. A cell fires if P is higher than a random number. The mobile entity's position is updated from a record of real motions and firing is again determined. Synapses between cells that both fire within 250 msec are strengthened. Synapses between cells with overlapping fields increase in strength whereas synapses between cells with distant fields will stay weak.

Abstract: A network system analyzes input words for the search and retrieval of pertinent information. The novel system then selects a module of a self organizing map (SOM) which contains the same dimension of classification as a selected input word and where neurons are interconnected horizontally (between modules) and vertically (at input and inside a module) by arrays of link weights. The system then determines what nodes or processing units within the SOM will be activated and subsequently compared to the selected input word. Feedback is utilized via a punishment/reward scheme to adjust the link weights so that the system learns the best paths and/or methods to create acceptable decisions or outputs.

Abstract: A signal compression system includes a coder and a decoder. The coder includes an extract unit for extracting an input feature vector from an input signal, a coder memory unit for storing a predesigned vector quantization (VQ) table for the coder such that the coder memory unit uses a set of primary indices to address entries within the pre-designed VQ table, a coder mapping unit for mapping indices from a set of secondary indices to the first set of indices, and a search unit for searching for one index out of the set of secondary indices, wherein the index from the set of secondary indices corresponds to an entry in the coder memory unit, and the entry best represents the input feature vector according to some predetermined criteria.

Abstract: A collection search system is responsive to a user query against a collection of documents to provide a search report. The collection search system includes a collection index including first predetermined single word and multiple word phrases as indexed terms occurring in the collection of documents, a linguistic parser that identifies a list of search terms from a user query, the linguistic parser identifying the list from second predetermined single words and multiple word phrases, and a search engine coupled to receive the list from the linguistic parser. The search engine operates to intersect the list with the collection index to identify a predetermined document from the collection of documents. The search engine includes an accumulator for summing a relevancy score for the predetermined document that is related to the intersection of the predetermined document with the list.

Abstract: A computer method of online mining of association rules by pre-processing data within the constraint of available memory. The required computational effort for performing online mining of association rules is reduced by a two-step procedure that involves first creating an adjacency lattice which pre-stores a number of large itemsets at a level of support dictated by available memory. The lattice structure is useful for both finding the itemsets quickly, by reducing the amount of disk I/O required to perform the analysis, and also using the itemsets in order to generate the rules. Once the adjacency lattice is obtained, the second (mining) step is further comprised of two phases. The first phase involves a search algorithm used to find the corresponding itemsets at user specified levels of minimum support. The second phase involves using those itemsets to generate association rules at the user specified level of minimum confidence.

Abstract: There is provided a network server which interfaces a client with selected database sites from a plurality of database sites. The network server comprises a meta-database (including both text information and multimedia information), a search agent, and a refining module. The search agent indexes the meta-database with a user query obtained from the client, and then distributes queries, developed pursuant to such indexing, to the selected ones of the plurality of database sites. In turn, database site information (responsive to the distributed queries) is retrieved from the selected ones of the plurality of database sites. A refining module is used to update the meta-database with the database relevancy information.

Abstract: The present invention relates to a system and methods for fine-granularity concurrency control in a parallel database. Very fine granularity (at the level of B-tree records) is implemented in a B-tree. Our method applies to B-trees, B-trees with variable keys, and their applications, such as semantic and object-oriented databases. Our method involves accumulating a transaction and then "optimistically" validating it, while attaining high efficiency with maximum semantic safety. "Lazy queries"--an efficient method for finding the intersection of two large queries--is provided for the system.

Abstract: An information managing device capable of systematically managing object data of more than one kind using a single database, having an object managing unit for storing the object data of more than one kind in general format, and a filter managing unit for storing a filter which specifies a specific format for each kind of object data. When the object data are registered, a control unit converts the object data the user has inputted in its specific format into those in general format to store the converted object data into the object managing unit. On the other hand, when the object data are retrieved, the control unit converts the object data in general format into those in specific format using the filter of the same kind as the desired object data.

Abstract: A search engine for retrieving documents pertinent to a query indexes documents in accordance with hyperlinks pointing to those documents. The indexer traverses the hypertext database and finds hypertext information including the address of the document the hyperlinks point to and the anchor text of each hyperlink. The information is stored in an inverted index file, which may also be used to calculate document link vectors for each hyperlink pointing to a particular document. When a query is entered, the search engine finds all document vectors for documents having the query terms in their anchor text. A query vector is also calculated, and the dot product of the query vector and each document link vector is calculated. The dot products relating to a particular document are summed to determine the relevance ranking for each document.

Abstract: A method enables a computer to manipulate data from a table in a remote database in accord with a query from a query requestor, wherein at least some data in the table includes large data objects (LOBs).

Abstract: A descriptive data structure provides an abstract representation of a rights management data structure such as a secure container. The abstract representation may describe, for example, the layout of the rights management data structure. It can also provide metadata describing or defining other characteristics of rights management data structure use and/or processing. For example, the descriptive data structure can provide integrity constraints that provide a way to state rules about associated information. The abstract representation can be used to create rights management data structures that are interoperable and compatible with one another. This arrangement preserves flexibility and ease of use without compromising security.

Abstract: In a multicast environment in which large-scale virtual environments are lt using locales and in which data from non-neighboring locales is ignored to minimize computation and local storage, a system is provided to find or track objects and their associated locales for locales to which a user is not listening by assigning the object a beacon and broadcasting the beacon in a subspace channel, with the beacon having the multicast address of the locale in which the sought-after object resides, thus to permit an upload of data from the particular locale. Beacons are doled out parsimoniously to objects one wants to find and objects one wants other people to find to minimize processing and data storage requirements. Ignored locales are thus located using a beacon broadcast in a subspace channel which in addition to broadcasting messages about itself to the locale it is in, also broadcasts messages into a subspace channel to others about itself and the multicast address of the locale it is in.

Abstract: A system, method, and apparatus for enabling a thin client to participate in a global transaction in a distributed object-oriented transaction processing environment that commits resources according to a two-phased commit protocol. The thin client is configured with a pseudo-transaction manager (PTM), and a remote server managing the commit procedure is configured with an object transaction manager having an interface to the (PTM). When an application on the thin client sends a request to begin a transaction, the PTM creates a temporary propagation context including a Global ID identifying the transaction and sets a flag in the context to indicate that this is a new transaction. The context is sent to a remote server for execution of a remote method.

Abstract: A system and method for retrieving the contents of a digital information system by multi-level categorization is disclosed. The system uses a file table, a category table, and a category bundling table. Headings and directions are displayed on navigation pages to guide the user to the desired file. The description and hierarchy relationship of categories, sub-categories, and files are stored in the tables. When the user selects a category or file and the direction of navigation, the system returns the file, or consults the tables and returns a list of new categories and files for a new navigation page. Moving down and then up through a category usually produces a different navigation page. Dynamic updating of the system categories and files can be easily accomplished by altering the content of the tables.

Abstract: An information server system is capable of optimization of multimedia data to be received from an information server depending upon transmission performance of a terminal for improving transmission efficiency and operability. The information server system is provided with an agent server as a gateway between a relatively low speed network, to which a terminal is connected, and a network connected to an information server, in which multimedia information forming one screen is divided into one or more image blocks and each block is transmitted therefrom with adding data amount and preference of thinning. The agent server is responsive to a demand for obtaining data from the terminal to the information server, determining thinning amount of data on the basis of the data amount and preference of thinning of respective blocks in the screen, added to the multimedia data from the information server depending upon transmission performance of the terminal.

Abstract: Disclosed is a process and apparatus for generating value lists. The operations for generating a value list include providing a first database having a plurality of fields configured to display values that are associated with predefined information types. Preferably a second database having a plurality of fields containing values associated with predefined information types is also provided. Once the databases are provided, one of the plurality of fields in the first database is selected to be associated with a value list. The selected field will preferably have a particular information type, and the value list will preferably contain one or more values that are specific to the particular information type. The process then proceeds to define the value list to contain values from certain fields of the second database.

Abstract: A data management system for file and database management including a design control system suitable for use in connection with the design of integrated circuits and other elements of manufacture having many parts which need to be developed in a concurrent engineering environment with inputs provided by users and or systems which may be located anywhere in the world providing a set of control information for coordinating movement of the design information through development and to release while providing dynamic tracking of the status of elements of the bills of materials in an integrated and coordinated activity control system utilizing a control repository which can be implemented in the form of a database (relational, object oriented, etc.) or using a flat file system.

Abstract: A method, apparatus, and computer program product for selecting and reviewing a distributed object installed on a distributed object system. A method of the invention includes generating a library of components corresponding to distributed objects on the distributed object system, which includes one component corresponding to the distributed object. Each of the components of the library includes information describing the distributed object to which the particular component corresponds. The contents of the library are displayed using a catalog interface device. The library is browsed using the catalog interface device to identify the component corresponding to the distributed object which is then selected. At least a portion of the information describing the distributed object is displayed.

Abstract: In an object-oriented database management system, a storage area of each data object is controlled using information indicating a storage area position of the data object in the database and a management serial number of the storage area. Each data object is provided with a directory object to keep therein information indicating a storage area position of the associated data object in the database and a management serial number of the storage area. In the database, the storage area of each data object is controlled using information denoting the storage area position in the corresponding directory object and the management serial number of the directory object.

Abstract: A multi-layer abstraction bucket mechanism connected between applications programs and at least one data source and providing to the users transformations of data and the results of processes performed on the data. The multi-layer abstraction bucket mechanism includes hierarchically connected abstraction layers, each including a methods object for storing methods for performing operations on data received from a data bucket of a hierarchically next lower abstraction layer, a data operation object for selecting a method to be executed by the method object, a data bucket for storing the results of an executed method, and a map for storing information for constructing the data bucket and for relating requests to methods residing in the methods object. The mechanism includes a data extraction layer and an abstraction layer. At least one abstraction layer is a data transformation layer while others include a data processing layer and a rules transformation layer for performing the rule based decision operations.

Abstract: A digital computer system in a clinical testing laboratory assays free Beta in a biological sample obtained from a pregnant patient and uses the resulting data, along with other patient and reference data, to create a patient profile. Operations of the system include inputting patient data for a pregnant patient into an electronic memory and creating an individual profile for each patient. The processor assigns specific accession numbers for each specimen, creates an assay format for a biological sample, and interfaces with the assay equipment to control the assaying of the biological sample for free Beta. The assay results are communicated into memory in machine readable form and used along with patient data and reference data to calculate a patient specific risk that the patient is carrying a fetus having a designated defect. From all the input data, the processor creates comprehensive patient profiles in both machine readable form and human readable form.

Abstract: A method and apparatus for managing access to resources is provided. When a process requires access to a resource, the process requests a lock on the resource from a lock manager unit that resides on the same node as the process. If a resource object for the resource does not exist, one is created in the lock manager unit, but not on lock manager units on other nodes. Because each lock manager unit does not have to store all resource objects, and resource objects are only created for resources that are actually used, the overhead of the lock management system is significantly reduced. Resources are grouped in recovery domains. When a lock manager unit that supported a recovery domain fails, the recovery domain is marked invalid. All resources in the recovery domain are considered invalid unless it would have been impossible for a failed instance to have held an exclusive lock on the resource. A snapshot of lock information is made before cleanup is performed on invalid resources.

Abstract: A design control system suitable for use in connection with the design of integrated circuits and other elements of manufacture having many parts which need to be developed in a concurrent engineering environment with inputs provided by users and or systems which may be located anywhere in the world providing a set of control information for coordinating movement of the design information through development and to release while providing dynamic tracking of the status of elements of the bills of materials in an integrated and coordinated activity control system utilizing a repository which can be implemented in the form of a database (relational, object oriented, etc.) or using a flat file system. Once a model is created and/or identified by control information design libraries hold the actual pieces of the design under control of the system without limit to the number of libraries, and providing for tracking and hierarchical designs which are allowed to traverse through multiple libraries.

Abstract: Information of a specific application, such as, CD-ROM or CD-DA is organized according to a specific information structure and also contains application data related thereto. There is disclosed a method of transferring information organized according to one or more information structures with an information signal subdivided into blocks. The header of each block contains an application code and application data, while the structure of the application data depends on the application code. According to a layered hierarchical structure the blocks are organized in groups having a group application code and the groups are organized in a volume having a volume application code. The invention makes it possible to know the information structure at any moment. Furthermore, an information carrier is described on which such an information signal is recorded. Also an arrangement for receiving and an arrangement for transmitting such information signals are described.

Abstract: When a log stream has a determined amount of reusable space, the log stream is compressed in order to free up the reusable space. Compression of the log stream includes rewriting one or more active entries located at one or more positions of the log stream to one or more other locations, either within the same log stream or on another log stream. Thereafter, the rewritten entries are deleted from their previous positions within the log stream. Additionally, one or more logically deleted entries of the log stream is removed from the log stream. The resulting log stream is a compressed log stream.

Abstract: One embodiment of the present invention presents a method and apparatus for efficiently performing garbage collection on objects defined within an object-oriented programming system. Garbage collection typically involves following pointers to determine which objects are presently being referenced so that other objects, that are not being referenced, can be removed. To this end, the present invention maintains a bitmap for each object that indicates which variables in the object are pointer variables and which variables are non-pointer variables. A garbage collection process examines the bitmap, and on the basis of the pattern contained in the bitmap jumps to a particular routine that is tailored to garbage collect the particular pattern of pointer and non-pointer values in the object. Note that the system includes a routine tailored for each possible bitmap pattern.

Abstract: A portable apparatus for directly acquiring text with a portable line scanner and organizing the acquired text in a selected format structure. The text is placed into the format structure as it is acquired. Control over the position of the acquired text in the format structure is provided by use of an ENTER key as well as other keys. Text attribute keys are provided for applying various attributes to the acquired text, such as italics, size, underlining, etc. Comments may be inserted by the user as desired and such comments are automatically placed inside a rectangular box and in a predetermined position. Footnotes having predefined text fields may also be inserted in the format structure by the user and the system displays a footnote panel to assist in creating the footnote.

Abstract: Embodiments of the present invention are capable of hiding a message, such as, for example, a copyright notice or other indicium of authorship, in an electronic document. An illustrative embodiment of the present invention comprises encoding an electronic document with a plurality of tags from a case insensitive markup language; and modulating the case of at least one tag character in at least one of said plurality of tags with a binary string, wherein the binary string represents a message to be hidden in the electronic document.

Abstract: The present invention relates to a document structure composing apparatus to judge a short part based upon the structure of the whole document, perform processing for complementation and compose document structure according to a desired document class. Complementation specification storage means stores the specification of complementation for document structure. Complementation means applies processing for complementation to pre-complemented document structure. Correlating rule storage means stores a rule for correlating components between different document classes. Document structure converting means converts an original document which meets the structural constraint of a specific document class to structure according to another document class according to the correlating rule. The original document is first converted from structure according to the specific document class to structure nearly according to a desired document class by the document structure converting means.

Abstract: The present invention includes a method of dynamic rendering protocol for display of glyphs used in a data processor comprising the steps of listing output image formats of font scalers which can be outputted to a graphics system and listing input image formats which can be inputted to a blitter function. These input image formats and output image formats are intersected, and an ordered list of the intersecting input image formats with output image formats are generated based upon a predetermined criteria. The first of the ordered list can then be selected for use.

Abstract: A computer bridge processes transactions in a computer system that includes a system memory. The bridge includes a first address decoder that allocates address space to the system memory according to a first allocation scheme and, in response to receiving transaction requests, directs the transaction requests to the system memory according to the first allocation scheme. The bridge also includes a second address decoder that allocates address space to the system memory and to a selected target device according to a second allocation scheme. In response to receiving transaction requests, the second address decoder directs the transactions requests to the system memory and the selected target device according to the second allocation scheme.

Abstract: A programmable circuit assembly and methods for high bandwidth data processing. The assembly includes an array of in-circuit programmable logic packages interconnected with an array of memory packages, allowing for elastic buffering of data in a variety of directions. Each programmable package includes package leads, a memory, and output drivers. Each output driver is connected to a respective package lead, which is configured to generate a logic function defined by programming data stored in the memory. A method includes storing programming data for operating the assembly to send signals on different paths between a programmable package and a memory package.

Abstract: A column address selection circuit 1 is provided, which renders all column address signals (CSi), from a start address to a stop address designated by a column address signal (ADCd+ADCu), to be a selective level when a segment address selection signal (SASj) and a block write signal (BW) are at an active level. A segment address selection circuit 2 is provided, which renders all segment address selection signals (SASj), from the start address to the stop address designated by the column address signal (ADCu) of a superordinate side, to be a selection level to supply it to the column address selection circuit 1, when the block write signal (BW) is at the active level.

Abstract: A non-volatile memory access protocol that facilitates concurrent accessing operations to multiple non-volatile memory components. This approach provides significant speed advantages over prior art non-volatile protocols. Also, power consumption is reduced in comparison to prior art synchronous protocols used for volatile memory because each memory component need not be continuously selected.

Abstract: A dynamic random access memory (DRAM) includes priority access control circuitry, where the DRAM has a first precharge mode and a priority precharge mode. In the first precharge mode, the array is precharged during an external row address strobe (RAS) and accessed during an active state of the external RAS. In the priority precharge mode, the array is precharged during a precharge state of an internal RAS initiated by a priority signal received during any one of the precharge and active states of the external RAS, where the priority access control circuitry returns to the first mode on a subsequent precharge state of the external RAS.

Abstract: A method and apparatus are provided for performing hierarchical address translation by translating each ternary hierarchical address into a binary address and a binary priority mask and storing the binary addresses in the binary CAM. A binary search of the priority masks is then performed by searching the CAM with a priority mask and choosing a next priority mask depending on the results of the search of the CAM until a correct matching entry (i.e., the matching entry with the lowest hierarchical level) is found. This technique only requires log.sub.2 N searches of the CAM, where N is the number of hierarchical levels represented by the priority field. A method and apparatus are also provided for performing hierarchical address translation by storing table entries including a priority field in a ternary CAM and performing only a fixed number of searches of the CAM.

Abstract: A method provides for waiting for a cache hit in the operation of a disk drive. The drive has an intelligent interface for communicating with a host, a magnetic disk, and a cache. The cache is divisible into a number of segments, and the cache employs a cache control structure. The drive receives a first command and a second command wherein the second command has a requested data range. The method comprises the steps of: (a) initializing a read in response to the first command to provide prefetch data to the cache in a prefetch data range wherein the prefetch data range has a first logical block address; and (b) waiting for a cache hit in response to the second command wherein the beginning of the requested data is within the prefetch data range and does not request the first logical block address.

Abstract: A cache memory automatically sets a low-, semi-, or high-speed mode operation according to a result of comparison between a half-period of a reference clock signal and a pulse width of a reference pulse signal provided by a reference pulse signal generator. Namely, a start signal generator generates a start signal used to access data memories, according to the frequency of the reference clock signal and a difference between the reference clock and pulse signals. According to the start signal and information indicating a hit tag memory, the cache memory dynamically switches the modes from one to another, without external instructions to the cache memory.

Abstract: An apparatus for processing a write miss signal from a copy-back data cache includes a load-store unit with an allocating load buffer, a non-allocating store buffer, and a priority control circuit to generate write-after-read hazards and read-after-write hazards to preserve the processing priority of entries within the allocating load buffer and the non-allocating store buffer. A prefetch circuit enqueues a prefetch command in the allocating load buffer and a store command in the non-allocating store buffer upon a write miss to the copy-back data cache. Thus, the priority control circuit forces a write-after-read hazard on the store command in the non-allocating store buffer. As a result, the prefetch command in the allocating load buffer secures an allocated line in the copy-back data cache, allowing the store command of the non-allocating store buffer to write data to the allocated line.

Abstract: A loop cache (26) is used in a data processing system for supplying instructions to a CPU to avoid accessing a main memory. Whether instructions stored in the loop cache can be supplied to the CPU is determined by a distributed TAG associated with the instruction address computed by the CPU. The instruction address includes an LCACHE index portion (42), an ITAG portion (44), and a GTAG (46). LCACHE index (42) selects corresponding locations in each of an ITAG array (50), an instruction array (52), and a valid bit array (54). A stored GTAG value (48) is chosen irrespective of where LCACHE index (42) is pointing. The GTAG portion of the instruction address (40) is compared to the stored GTAG value (48). The ITAG portion (44) of instruction address (40) is compared with the indexed ITAG of the ITAG array (50). If both the GTAG and ITAG compare favorably, the instruction is supplied from the loop cache to the CPU, rather than from main memory.