Patent application number | Description | Published |
20100058346 | Assigning Threads and Data of Computer Program within Processor Having Hardware Locality Groups - A computer program having threads and data is assigned to a processor having a processor cores and memory organized over hardware locality groups. The computer program is profiled to generate a data thread interaction graph (DTIG) representing the computer program. The threads and the data of the computer program are organized over clusters using the DTIG and based on one or more constraints. The DTIG is displayed to a user, and the user is permitted to modify the constraints such that the threads and the data of the computer program are reorganized over the clusters. Each cluster is mapped onto one of the hardware locality groups. The computer program is regenerated based on the mappings of clusters to hardware locality groups. At run-time, optimizations are performed to improve execution performance, while the computer program is executed. | 03-04-2010 |
20100094870 | METHOD FOR MASSIVELY PARALLEL MULTI-CORE TEXT INDEXING - There is provided, in a parallel pipelined structure on a multi-core device, a method for parallel pipelined multi-core indexing. The method includes generating one or more single document indexes respectively corresponding to one or more single documents of a given data stream. The method further includes generating one or more multi-document interval-based hash tables from the one or more single document indexes. The method also includes generating a global hash table formed from merging one or more of the multi-document interval-based hash tables, the global hash table representing a collective index for all of the single documents for which the one or more single document indexes were generated. | 04-15-2010 |
20110099553 | SYSTEMS AND METHODS FOR AFFINITY DRIVEN DISTRIBUTED SCHEDULING OF PARALLEL COMPUTATIONS - Embodiments of the invention provide efficient scheduling of parallel computations for higher productivity and performance. Embodiments of the invention provide various methods effective for affinity driven and distributed scheduling of multi-place parallel computations with physical deadlock freedom. | 04-28-2011 |
20110252033 | SYSTEM AND METHOD FOR MULTITHREADED TEXT INDEXING FOR NEXT GENERATION MULTI-CORE ARCHITECTURES - A system and method for indexing documents in a data storage system includes generating a single document hash table in storage memory for a single document using an index construction in a multithreaded and scalable configuration wherein multiple threads are each assigned work to reduce synchronization between threads. The single document hash table includes partitioning the single document and indexing strings of partitioned portions of the single document to create a minor hash table for each document sub-part; generating a document level hash table from the minor hash tables; updating a stream level hash table for the strings which maps every string to a global identifier; and generating a term reordered array from the document level hash table. | 10-13-2011 |
20120102003 | PARALLEL DATA REDUNDANCY REMOVAL - A method, system, and computer usable program product for parallel data redundancy removal are provided in the illustrative embodiments. A plurality of values is computed for a record in a plurality of records stored in a storage device. The plurality of values for the record is distributed to corresponding queues in a plurality of queues, wherein each of the plurality of queues is associated with a corresponding section of a Bloom filter. A determination is made whether each value distributed to the corresponding queues for the record is indicated by a corresponding value in the corresponding section of the Bloom filter. The record is identified as a redundant record in response to a determination that each value distributed to the corresponding queues for the record is indicated by a corresponding value in the corresponding section of the Bloom filter. | 04-26-2012 |
20120142319 | SYSTEMS AND METHODS FOR JOINT ANALYTICS ON USER LEVEL AND NETWORK LEVEL DATA OF A COMMUNICATIONS NETWORK - Systems and associated methods provide for joint analytics on user level data and network level data. Systems and methods provide for accessing telecommunication network user level data and network level data and integrating analysis on both data types to produce a common result. Embodiments utilize joint analytics to generate patterns and rules concerning the content and services accessed by a user, when they are accessed, and how they are accessed. | 06-07-2012 |
20120166728 | SYSTEMS AND METHODS FOR PERFORMING PARALLEL MULTI-LEVEL DATA COMPUTATIONS - Systems and methods for performing parallel multi-level data computations in a storage system are provided. One system includes a memory storing data, multiple caches, and a processor. The processor is configured to perform the method below. One method includes determining the total amount of data in the memory, dividing the amount of data by each cache capacity to determine the number of nodes needed for processing the data in the memory, and automatically creating the nodes. Here, the nodes form a tree structure including multiple levels where the lowest level includes a first number of nodes equal to the amount of data divided by the cache memory capacity. Also, each lowest level node is configured to process an amount of data equal to the cache memory capacity and each level above the lowest level is configured to include one or more nodes for receiving an input from lower level nodes. | 06-28-2012 |
20130086355 | Distributed Data Scalable Adaptive Map-Reduce Framework - A method, an apparatus and an article of manufacture for generating a distributed data scalable adaptive map-reduce framework for at least one multi-core cluster. The method includes partitioning a cluster into at least one computational group, determining at least one key-group leader within each computational group, performing a local combine operation at each computational group, performing a global combine operation at each of the at least one key-group leader within each computational group based on a result from the local combine operation, and performing a global map-reduce operation across the at least one key-group leader within each computational group. | 04-04-2013 |
20130086356 | Distributed Data Scalable Adaptive Map-Reduce Framework - A method for generating a distributed data scalable adaptive map-reduce framework for at least one multi-core cluster. The method includes partitioning a cluster into at least one computational group, determining at least one key-group leader within each computational group, performing a local combine operation at each computational group, performing a global combine operation at each of the at least one key-group leader within each computational group based on a result from the local combine operation, and performing a global map-reduce operation across the at least one key-group leader within each computational group. | 04-04-2013 |
20130172043 | ONLINE AND DISTRIBUTED OPTIMIZATION FRAMEWORK FOR WIRELESS ANALYTICS - A method, computer program product, and computer system directed to an online and distributed optimization framework for wireless analytics. A radio network controller determines a ranking for each of a plurality of received objects using a plurality of similarity graphs. The radio network controller extracts a common structure by collaborative filtering data associated with a plurality of user devices and the plurality of received objects. The common structure is analyzed to infer usage patterns within a time slot. The radio network controller stores a subset of the ranked objects of the plurality of received objects in response to the analysis. | 07-04-2013 |
20140067808 | Distributed Scalable Clustering and Community Detection - Techniques, an apparatus and an article of manufacture for distributed scalable clustering and community detection. A method includes generating a label for each node in a graph, wherein said label identifies a community in which a node participates, propagating each label locally within two or more segments of the graph based on a participation percentage of each node in at least one identified community within the graph, and deriving at least one cluster of nodes in the graph that corresponds to the at least one identified community based on said propagating. | 03-06-2014 |