Hewlett-packard Development Company, L.p.

Hewlett Packard Enterprise Development LP

Entit Software LLC

Indrajit Roy

Vanish Talwar

Mehul A Shah

Robert S Schreiber

Nathan Lorenzo Binkert

Alvin AuYoung

Parthasarathy Ranganathan

Wojciech Golab

Shivaram Venkataraman

Michael Mihn-Jong Lee

Meichun Hsu

Kimberly Keeton

Yuan Chen

Krishnamurthy Viswanathan

Benjamin Albert Sowell

Stavros Harizopoulos

Haris Volos

Shreya Prasad

Edward Ma

Vishrut Gupta

Vincent Xu

Sunil Venkayala

Dan Huang

Kyungyong Lee

Henggang Cui

Erik Tamas Bodzsar

Naila Farooqui

Sangman Kim

A transformation on raw data is applied to produce transformed data, where the transformation includes at least one selected from among a summary of the raw data or a transform of the raw data between different domains. In response to a query to access data, the query is processed using the transformed data.

Processing a query using transformed raw data

According to an example, an instruction to run a kernel of an application on an apparatus having a first processing unit integrated with a second processing unit may be received. In addition, an application profile for the application at a runtime of the application kernel on the second processing unit may be created, in which the application profile identifies an affinity of the application kernel to be run on either the first processing unit or the second processing unit, and identifies a characterization of an input data set of the application. The application profile may also be stored in a data store.

Integrated heterogeneous processing units

The communication between integrated graphics processing units (GPUs) is disclosed. A first integrated GPU of a first computing device obtains a tuple pertaining to data to be transmitted to a second integrated GPU of a second computing device. The tuple comprises at least a length of the data. The first integrated GPU allocates a virtual address space to the data based on the length of the data, where the virtual address space has a plurality of virtual addresses. Further, a mapping table of a mapping between the plurality of virtual addresses and a plurality of bus addresses is provided by the first integrated GPU to a communication module of the first computing device to transmit the data, where the plurality of bus addresses indicate physical locations of the data.

Communication between integrated graphics processing units

Examples disclosed herein relate to distributed computing in R. Some examples disclosed herein may include identifying a distributed multivariate apply (dmapply) operation and an invocation of a distributed computing backend and determining a function referenced in the dmapply operation. A distributed backend driver associated with the invoked distributed computing backend may translate the determined function to a function native to an R application programming interface (API) of the invoked distributed computing backend and may provide the translated function to the invoked distributed computing backend to perform the translated function on a distributed data set referenced in the dmapply operation.

Distributed computing in R

A method, system, and computer-readable storage device for transferring data from a database management system to a data analytics system is described herein. In one aspect, a database management node storing a data partition of a database table that is distributed across a plurality of database management nodes of a database management system receives a data transfer request from a data analytics node of a data analytics node cluster. The data transfer request may include a data transfer policy for splitting the data partition onto the data analytics node cluster. A data split may be identified from the data partition. The data split may be a subset of the data partition. The data split is communicated to the data analytics node according to the data transfer policy.

Data transfer requests with data transfer policies

An embodiment provides a system and method for transaction commitment and replication. The method includes receiving a minitransaction from a client node at one or more memory nodes, wherein each memory node includes a number of replicas. The minitransaction is a type of transaction which atomically executes any combination of reading, comparing, and writing to any of a number of memory locations. The method also includes determining, for a leader of the replicas within a memory node, whether the leader is able to commit the minitransaction and stabilizing state changes of the minitransaction within a transaction log using a consensus procedure to update the replicas. The method further includes committing the minitransaction if, at each memory node, a quorum of the replicas is able to stabilize the minitransaction, or aborting the minitransaction otherwise.

Transaction commitment and replication in a storage system

A sparse array is partitioned into first partitions and a second array is partitioned into second partitions based on an invariant relationship between the sparse array and the second array. The sparse array and the second array are associated with a computation involving the sparse array and the second array. The first partitions and the corresponding second partitions are distributed to workers. A different first partition and corresponding second partition is distributed to each of the workers. Third partitions of the sparse array and corresponding fourth partitions of the second array are determined based on the invariant relationship and measurements of load are received from each of the workers. At least one of the first partitions and the corresponding second partition is different from one of the third partitions and the corresponding fourth partition. The at least one of the first partitions and the corresponding second partition that is different is redistributed among the workers. A different third partition and corresponding fourth partition is executed by each of the workers.

Load-balanced sparse array processing

A technique includes distributing a plurality of tasks among a plurality of worker nodes to perform a processing operation on an array. Each task is associated with a set of a least one data block of the array, and an order of the tasks is defined by an array-based programming language. Distribution of the tasks includes, for at least one of the worker nodes, selectively reordering the order defined by the array-based programming language to regulate an amount of data transferred to the worker node.

Distributing processing of array block tasks

A plurality of array partitions are defined for use by a set of tasks of the program run-time. The array partitions can be determined from one or more arrays that are utilized by the program at run-time. Each of the plurality of computing devices are assigned to perform one or more tasks in the set of tasks. By assigning each of the plurality of computing devices to perform one or more tasks, an objective to reduce data transfer amongst the plurality of computing devices can be implemented.

Managing array computations during programmatic run-time in a distributed computing environment

In a method of implementing a graph storage system, the graph storage system is stored on a plurality of computing systems. A global address space is provided for distributed graph storage. The global address space is managed with graph allocators, in which a graph allocator allocates space from a block of the distributed global memory in order to store a plurality of graphs.

Distributed graph storage system

A distributed balanced tree having a plurality of nodes distributed across a plurality of servers is accessed. Version information associated with a node of the distributed balanced tree is retrieved. Validity of a lookup transaction performed on the balanced tree is determined by verifying a version value of only the leaf node accessed during the lookup operation against the retrieved version information.

Validation of distributed balanced trees

Examples of the present disclosure may include methods, systems, and computer readable media with executable instructions. An example method for event processing for graph-structured data can include storing graph structured data. The graph structured data includes a plurality of vertex, edge, and/or property graph elements. The example method further includes defining a first graph view of a characteristic of vertex, edge, and/or property graph elements. A subgraph is determined as a subset of the plurality of vertex, edge, and/or property graph elements that have the characteristic of vertex, edge, and/or property graph elements defined by the first graph view. The vertex, edge, and/or property graph elements of the subgraph are processed responsive to a predefined event that occurs on at least one of the vertex, edge, and/or property graph elements of the subgraph.

Growing community of inventors

Mountain View, CA, United States of America

Indrajit Roy