-
公开(公告)号：US10228922B2
公开(公告)日：2019-03-12
申请号：US14993722
申请日：2016-01-12
Applicant: INTERNATIONAL BUSINESS MACHINES CORPORATION
Inventor： Matthias Boehm , Douglas Burdick , Berthold Reinwald , Prithviraj Sen , Shirish Tatikonda , Yuanyuan Tian , Shivakumar Vaithyanathan
Abstract: Parallel execution of machine learning programs is provided. Program code is received. The program code contains at least one parallel for statement having a plurality of iterations. A parallel execution plan is determined for the program code. According to the parallel execution plan, the plurality of iterations is partitioned into a plurality of tasks. Each task comprises at least one iteration. The iterations of each task are independent. Data required by the plurality of tasks is determined. An access pattern by the plurality of tasks of the data is determined. The data is partitioned based on the access pattern.
-
2.发明申请HYBRID PARALLELIZATION STRATEGIES FOR MACHINE LEARNING PROGRAMS ON TOP OF MAPREDUCE 审中-公开用于MAPREDUCE顶部机器学习程序的混合并行策略公开(公告)号：US20160124730A1
公开(公告)日：2016-05-05
申请号：US14993722
申请日：2016-01-12
Applicant: INTERNATIONAL BUSINESS MACHINES CORPORATION
Inventor： Matthias Boehm , Doughlas Burdick , Berthold Reinwald , Prithviraj Sen , Shirish Tatikonda , Yuanyuan Tian , Shivakumar Vaithyanathan
CPC classification number: G06F8/445 , G06F8/443 , G06F8/45 , G06F8/452 , G06F9/4881
Abstract: Parallel execution of machine learning programs is provided. Program code is received. The program code contains at least one parallel for statement having a plurality of iterations. A parallel execution plan is determined for the program code. According to the parallel execution plan, the plurality of iterations is partitioned into a plurality of tasks. Each task comprises at least one iteration. The iterations of each task are independent. Data required by the plurality of tasks is determined. An access pattern by the plurality of tasks of the data is determined. The data is partitioned based on the access pattern.
Abstract translation: 提供机器学习程序的并行执行。 接收到程序代码。 程序代码包含至少一个具有多个迭代的并行的语句。 确定程序代码的并行执行计划。 根据并行执行方案，将多个迭代划分为多个任务。 每个任务包括至少一个迭代。 每个任务的迭代是独立的。 确定多个任务所需的数据。 确定数据的多个任务的访问模式。 数据根据访问模式进行分区。
-
3.发明申请HYBRID PARALLELIZATION STRATEGIES FOR MACHINE LEARNING PROGRAMS ON TOP OF MAPREDUCE 有权用于MAPREDUCE顶部机器学习程序的混合并行策略公开(公告)号：US20150378696A1
公开(公告)日：2015-12-31
申请号：US14317016
申请日：2014-06-27
Applicant: INTERNATIONAL BUSINESS MACHINES CORPORATION
Inventor： Matthias Boehm , Douglas Burdick , Berthold Reinwald , Prithviraj Sen , Shirish Tatikonda , Yuanyuan Tian , Shivakumar Vaithyanathan
IPC: G06F9/45
CPC classification number: G06F8/445 , G06F8/443 , G06F8/45 , G06F8/452 , G06F9/4881
Abstract: Hybrid parallelization strategies for machine learning programs on top of MapReduce are provided. In one embodiment, a method of and computer program product for parallel execution of machine learning programs are provided. Program code is received. The program code contains at least one parallel for statement having a plurality of iterations. A parallel execution plan is determined for the program code. According to the parallel execution plan, the plurality of iterations is partitioned into a plurality of tasks. Each task comprises at least one iteration. The iterations of each task are independent.
Abstract translation: 提供了MapReduce之上的机器学习程序的混合并行化策略。 在一个实施例中，提供了用于并行执行机器学习程序的方法和计算机程序产品。 接收到程序代码。 程序代码包含至少一个具有多个迭代的并行的语句。 确定程序代码的并行执行计划。 根据并行执行方案，将多个迭代划分为多个任务。 每个任务包括至少一个迭代。 每个任务的迭代是独立的。
-
公开(公告)号：US09684493B2
公开(公告)日：2017-06-20
申请号：US14293223
申请日：2014-06-02
Applicant: International Business Machines Corporation
Inventor： Matthias Boehm , Douglas R. Burdick , Stefan Burnicki , Berthold Reinwald , Shirish Tatikonda
CPC classification number: G06F8/41
Abstract: In a method for analyzing a large data set using a statistical computing environment language operation, a processor generates code from the statistical computing environment language operation that can be understood by a software system for processing machine learning algorithms in a MapReduce environment. A processor transfers the code to the software system for processing machine learning algorithms in a MapReduce environment. A processor invokes execution of the code with the software system for processing machine learning algorithms in a MapReduce environment.
-
公开(公告)号：US20170147943A1
公开(公告)日：2017-05-25
申请号：US14949740
申请日：2015-11-23
Applicant: International Business Machines Corporation
Inventor： Matthias Boehm , Mathias Peters , Berthold Reinwald , Shirish Tatikonda
IPC: G06N99/00
Abstract: A method for global data flow optimization for machine learning (ML) programs. The method includes receiving, by a storage device, an initial plan for an ML program. A processor builds a nested global data flow graph representation using the initial plan. Operator directed acyclic graphs (DAGs) are connected using crossblock operators according to inter-block data dependencies. The initial plan for the ML program is re-written resulting in an optimized plan for the ML program with respect to its global data flow properties. The re-writing includes re-writes of: configuration dataflow properties, operator selection and structural changes.
-
Patent Agency Ranking
公开(公告)号：US20160364327A1
公开(公告)日：2016-12-15
申请号：US15252714
申请日：2016-08-31
Applicant: International Business Machines Corporation
Inventor： Berthold Reinwald , Shirish Tatikonda , Yuanyuan Tian
CPC classification number: G06F12/0223 , G06F17/16 , G06F2212/251
Abstract: Embodiments of the invention relate to sparsity-driven matrix representation. In one embodiment, a sparsity of a matrix is determined and the sparsity is compared to a threshold. Computer memory is allocated to store the matrix in a first data structure format based on the sparsity being greater than the threshold. Computer memory is allocated to store the matrix in a second data structure format based on the sparsity not being greater than the threshold
-
公开(公告)号：US20150347101A1
公开(公告)日：2015-12-03
申请号：US14293223
申请日：2014-06-02
Applicant: International Business Machines Corporation
Inventor： Matthias Boehm , Douglas R. Burdick , Stefan Burnicki , Berthold Reinwald , Shirish Tatikonda
IPC: G06F9/45
CPC classification number: G06F8/41
Abstract: In a method for analyzing a large data set using a statistical computing environment language operation, a processor generates code from the statistical computing environment language operation that can be understood by a software system for processing machine learning algorithms in a MapReduce environment. A processor transfers the code to the software system for processing machine learning algorithms in a MapReduce environment. A processor invokes execution of the code with the software system for processing machine learning algorithms in a MapReduce environment.
Abstract translation: 在使用统计计算环境语言操作分析大数据集的方法中，处理器从统计计算环境语言操作生成可由MapReduce环境中用于处理机器学习算法的软件系统理解的代码。 处理器将代码传输到软件系统，以便在MapReduce环境中处理机器学习算法。 处理器在MapReduce环境中调用用于处理机器学习算法的软件系统的代码执行。
-
公开(公告)号：US10198291B2
公开(公告)日：2019-02-05
申请号：US15452571
申请日：2017-03-07
Applicant: INTERNATIONAL BUSINESS MACHINES CORPORATION
Inventor： Matthias Boehm , Berthold Reinwald , Shirish Tatikonda
Abstract: One embodiment provides a method for runtime piggybacking of concurrent data-parallel jobs in task-parallel machine learning (ML) programs including intercepting, by a processor, executable jobs including executable map reduce (MR) jobs and looped jobs in a job stream. The processor queues the executable jobs, and applies runtime piggybacking of multiple jobs by processing workers of different types. Runtime piggybacking for a ParFOR (parallel for) ML program is optimized including configuring the runtime piggybacking based on processing worker type, degree of parallelism and minimum time thresholds.
-
9.发明申请公开(公告)号：US20180211357A1
公开(公告)日：2018-07-26
申请号：US15924029
申请日：2018-03-16
Applicant: International Business Machines Corporation
Inventor： Arash Ashari , Matthias Boehm , Keith W. Campbell , Alexandre Evfimievski , John D. Keenleyside , Berthold Reinwald , Shirish Tatikonda
IPC: G06T1/20
CPC classification number: G06T1/20
Abstract: A method for optimization of machine learning (ML) workloads on a graphics processor unit (GPU). The method includes identifying a computation having a generic pattern commonly observed in ML processes. Hierarchical aggregation spanning a memory hierarchy of the GPU for processing is performed for the identified computation including maintaining partial output vector results in shared memory of the GPU. Hierarchical aggregation for vectors is performed including performing intra-block aggregation for multiple thread blocks of a partial output vector results on GPU global memory.
-
公开(公告)号：US09972063B2
公开(公告)日：2018-05-15
申请号：US14813522
申请日：2015-07-30
Applicant: INTERNATIONAL BUSINESS MACHINES CORPORATION
Inventor： Arash Ashari , Matthias Boehm , Keith W. Campbell , Alexandre Evfimievski , John D. Keenleyside , Berthold Reinwald , Shirish Tatikonda
IPC: G06T1/20
CPC classification number: G06T1/20
Abstract: A method for optimization of machine learning (ML) workloads on a graphics processor unit (GPU). The method includes identifying a computation having a generic pattern commonly observed in ML processes. An optimized fused GPU kernel is employed to exploit temporal locality for inherent data-flow dependencies in the identified computation. Hierarchical aggregation spanning a memory hierarchy of the GPU for processing for the identified computation is performed. GPU kernel launch parameters are estimated following an analytical model that maximizes thread occupancy and minimizes atomic writes to GPU global memory.
-
-
-
-
-
-
-
-
-
Search Results
26
records
(took 0.023 seconds)
