专利快速检索-快速检索全球专利，免费商用专利数据库-IPRDB

1. 发明公开

EP3360054A1 SYSTEMS AND METHODS FOR CREATING AND USING A DATA STRUCTURE FOR PARALLEL PROGRAMMING 有权
公开(公告)号：EP3360054A1
公开(公告)日：2018-08-15
申请号：EP16861427.9
申请日：2016-10-18
申请人： Huawei Technologies Co., Ltd.
发明人： CHEN, Liya , TIAN, Chen , HU, Ziang
IPC分类号： G06F17/00
CPC分类号： G06F15/17318 , G06F8/314 , G06F9/30036 , G06F9/3009 , G06F9/445 , G06F9/50
摘要： System and method embodiments are provided for creating data structure for parallel programming. A method for creating data structures for parallel programming includes forming, by one or more processors, one or more data structures, each data structure comprising one or more global containers and a plurality of local containers. Each of the global containers is accessible by all of a plurality of threads in a multi-thread parallel processing environment. Each of the plurality of local containers is accessible only by a corresponding one of the plurality of threads. A global container is split into a second plurality of local containers when items are going to be processed in parallel and two or more local containers are merged into a single global container when a parallel process reaches a synchronization point.

2. 发明公开

EP3407268A1 DEVICE AND METHOD FOR EXECUTING REVERSAL TRAINING OF ARTIFICIAL NEURAL NETWORK 审中-公开
公开(公告)号：EP3407268A1
公开(公告)日：2018-11-28
申请号：EP16885905.6
申请日：2016-04-01
申请人： Cambricon Technologies Corporation Limited
发明人： LIU, Shaoli , GUO, Qi , CHEN, Yunji , CHEN, Tianshi
IPC分类号： G06N3/08
CPC分类号： G06N3/084 , G06F9/22 , G06F9/30145 , G06F9/3838 , G06F13/28 , G06F15/17318 , G06N3/0454 , G06N3/08
摘要： An apparatus for executing backwards training of an artificial neural network comprises an instruction caching unit 1, a controller unit 2, a direct memory access unit 3, an H tree module 4, a master computation module 5, and multiple slave computation modules 6. By using the apparatus, backwards training of a multiple-layer artificial neural network can be implemented. For each layer, firstly, weighted summation is performed on input gradient vectors to calculate an output gradient vector of this layer. The output gradient vector is multiplied by a derivative value of a next-layer activation function on which forward operation is performed, so that a next-layer input gradient vector can be obtained. The input gradient vector is multiplied by an input neuron counterpoint in forward operation to obtain the gradient of a weight value of this layer, and the weight value of this layer can be updated according to the gradient of the obtained weight value of this layer.

3. 发明公开

EP2416255A4 DATA TRANSFER CONTROL DEVICE, SYSTEM, AND METHOD 审中-公开
标题翻译：数据传输控制装置，系统及方法
公开(公告)号：EP2416255A4
公开(公告)日：2012-12-05
申请号：EP09842911
申请日：2009-03-30
申请人： FUJITSU LTD
发明人： KAWAHARA JUN
IPC分类号： G06F15/173 , G06F13/36
CPC分类号： G06F9/54 , G06F13/4022 , G06F15/17318

4. 发明公开

EP2984570A4 MULTIPROCESSOR SYSTEM WITH INDEPENDENT DIRECT ACCESS TO BULK SOLID STATE MEMORY RESOURCES 审中-公开
公开(公告)号：EP2984570A4
公开(公告)日：2017-11-08
申请号：EP14783396
申请日：2014-04-09
申请人： EMC CORP
发明人： CARLSON FREDERIC ROY JR , HIMELSTEIN MARK , WILFORD BRUCE , ARAI DAN , EMBERSON DAVID R
IPC分类号： G06F12/00
CPC分类号： G11C7/1072 , G06F1/00 , G06F12/08 , G06F15/167 , G06F15/17318 , G06F17/30327

5. 发明公开

EP2972912A4 A METHOD FOR IMPLEMENTING A LINE SPEED INTERCONNECT STRUCTURE 审中-公开
标题翻译： VERFAHREN ZUR IMPLEMENTIERUNG EINER ZEILengesCHWINDIGKEITSVERBINDUNGSSTRUKTUR
公开(公告)号：EP2972912A4
公开(公告)日：2016-11-09
申请号：EP14768040
申请日：2014-03-13
申请人： SOFT MACHINES INC
发明人： ABDALLAH MOHAMMAD
IPC分类号： G06F13/14 , G06F9/38 , G06F9/46 , G06F12/00 , G06F13/00 , G06F13/38 , G06F15/173 , H04L12/28 , H04L12/933
CPC分类号： G06F9/3885 , G06F9/3004 , G06F9/3853 , G06F9/462 , G06F12/00 , G06F12/0853 , G06F12/0886 , G06F13/00 , G06F15/17318 , G06F2212/1041 , G06F2212/1056 , H04L12/28 , H04L49/1523
摘要： A method for line speed interconnect processing. The method includes receiving initial inputs from an input communications path, performing a pre-sorting of the initial inputs by using a first stage interconnect parallel processor to create intermediate inputs, and performing the final combining and splitting of the intermediate inputs by using a second stage interconnect parallel processor to create resulting outputs. The method further includes transmitting the resulting outputs out of the second stage at line speed.
摘要翻译：一种线速互连处理方法。该方法包括从输入通信路径接收初始输入，通过使用第一级互连并行处理器来执行初始输入的预排序以产生中间输入，以及通过使用第二级执行中间输入的最终组合和分割互连并行处理器来创建结果输出。该方法还包括以线速度将所得到的输出从第二级传输。

6. 发明公开

EP2860643A2 Collective communications apparatus and method for parallel systems 审中-公开
标题翻译： Kollektive Kommunikationsvorrichtung und Verfahrenfürparallele Systeme
公开(公告)号：EP2860643A2
公开(公告)日：2015-04-15
申请号：EP14186688.9
申请日：2014-09-26
申请人： Intel Corporation
发明人： Knies, Allan D , Keppel, David Pardo , Woo, Dong Hyuk , Fryman, Joshua B
IPC分类号： G06F15/173 , G06F9/52
CPC分类号： G06F13/4068 , G06F9/52 , G06F15/17318 , G06F15/17325
摘要： A collective communication apparatus and method for parallel computing systems. For example, one embodiment of an apparatus comprises a plurality of processor elements (PEs); collective interconnect logic to dynamically form a virtual collective interconnect (VCI) between the PEs at runtime without global communication among all of the PEs, the VCI defining a logical topology between the PEs in which each PE is directly communicatively coupled to a only a subset of the remaining PEs; and execution logic to execute collective operations across the PEs, wherein one or more of the PEs receive first results from a first portion of the subset of the remaining PEs, perform a portion of the collective operations, and provide second results to a second portion of the subset of the remaining PEs.
摘要翻译：一种用于并行计算系统的集体通信装置和方法。例如，设备的一个实施例包括多个处理器元件（PE）; 集体互连逻辑以在运行时动态地形成PE之间的虚拟集体互连（VCI）而不在所有PE之间进行全局通信，VCI在PE之间定义逻辑拓扑，其中每个PE直接通信地耦合到仅一个子集余下的PE; 以及用于在所述PE之间执行集合操作的执行逻辑，其中所述PE中的一个或多个从所述剩余PE的子集的第一部分接收第一结果，执行所述集体操作的一部分，并且将第二结果提供给其余PE的子集。

7. 发明公开

EP2757748A1 Information processing system and method of controlling the same 有权
标题翻译： Informationsverarbeitungssystem und Steuerverfahrendafür
公开(公告)号：EP2757748A1
公开(公告)日：2014-07-23
申请号：EP13193949.8
申请日：2013-11-21
申请人： FUJITSU LIMITED
发明人： Nose, Takafumi , Miura, Kenichi
IPC分类号： H04L12/721 , G06F15/173
CPC分类号： G06F3/0655 , G06F3/061 , G06F3/0671 , G06F15/17318 , H04L45/06
摘要： An information processing system in which a plurality of information processing apparatuses are connected with each other, wherein each information processing apparatus includes a storage unit configured to store data according to each destination information processing apparatus, and a transmission control unit configured to transmit data to be transmitted in the same transmission direction and with the same number of hops collectively among the data stored in the storage unit.
摘要翻译：一种信息处理系统，其中多个信息处理设备彼此连接，其中每个信息处理设备包括：存储单元，被配置为根据每个目的地信息处理设备存储数据;以及发送控制单元，被配置为将数据发送到在存储在存储单元中的数据之间以相同的传输方向和相同数量的跳数共同发送。

8. 发明公开

EP2416255A1 DATA TRANSFER CONTROL DEVICE, SYSTEM, AND METHOD 审中-公开
标题翻译：数据传输控制设备，系统和方法
公开(公告)号：EP2416255A1
公开(公告)日：2012-02-08
申请号：EP09842911.1
申请日：2009-03-30
申请人： Fujitsu Limited
发明人： KAWAHARA, Jun
IPC分类号： G06F15/173 , G06F13/36
CPC分类号： G06F9/54 , G06F13/4022 , G06F15/17318
摘要： A data transfer controlling apparatus includes a first controlling unit that performs a broadcast transfer to a plurality of nodes connected to a first route switching device, and requests a second node representing nodes connected to a second route switching device that is connected to the first route switching device to perform a broadcast transfer when the first controlling unit receives a request of the broadcast transfer from a first node connected to the first route switching device, and a second controlling unit that performs the broadcast transfer to all of nodes connected to the first route switching device when the second controlling unit receives a request of a broadcast transfer from the second node.
摘要翻译：一种数据传送控制设备包括：第一控制单元，其向连接到第一路由交换设备的多个节点执行广播传送，并且请求表示连接到与第一路由交换设备连接的第二路由交换设备的节点的第二节点设备，用于当第一控制单元从连接到第一路由交换设备的第一节点接收到广播传送的请求时执行广播传送;以及第二控制单元，用于执行广播传送到连接到第一路由交换的所有节点设备，当第二控制单元从第二节点接收到广播传送的请求时。

9. 发明授权

EP3877863B1 APPARATUS, METHOD AND COMPUTER PROGRAM PRODUCT FOR PERFORMING A COLLECTIVE COMMUNICATION OPERATION IN A DATA COMMUNICATIONS NETWORK 有权
公开(公告)号：EP3877863B1
公开(公告)日：2024-04-24
申请号：EP18845422.7
申请日：2018-12-13
IPC分类号： G06F15/173
CPC分类号： G06F15/17318

10. 发明公开

EP2860643A3 Collective communications apparatus and method for parallel systems 审中-公开
标题翻译：用于并行系统集体通信设备和方法
公开(公告)号：EP2860643A3
公开(公告)日：2016-05-04
申请号：EP14186688.9
申请日：2014-09-26
申请人： Intel Corporation
发明人： Knies, Allan D , Keppel, David Pardo , Woo, Dong Hyuk , Fryman, Joshua B
IPC分类号： G06F15/173 , G06F9/52 , G06F13/40
CPC分类号： G06F13/4068 , G06F9/52 , G06F15/17318 , G06F15/17325
摘要： A collective communication apparatus and method for parallel computing systems. For example, one embodiment of an apparatus comprises a plurality of processor elements (PEs); collective interconnect logic to dynamically form a virtual collective interconnect (VCI) between the PEs at runtime without global communication among all of the PEs, the VCI defining a logical topology between the PEs in which each PE is directly communicatively coupled to a only a subset of the remaining PEs; and execution logic to execute collective operations across the PEs, wherein one or more of the PEs receive first results from a first portion of the subset of the remaining PEs, perform a portion of the collective operations, and provide second results to a second portion of the subset of the remaining PEs.

你已经成功收藏专利！

检索式保存成功!

IPRDB

热门服务

关于我们

友情链接

联系方式