公开(公告)号：US20160011996A1

公开(公告)日：2016-01-14

申请号：US14701371

申请日：2015-04-30

Applicant: International Business Machines Corporation

Inventor： Sameh Asaad ,  Ralph E. Bellofatto ,  Michael A. Blocksome ,  Matthias A. Blumrich ,  Peter Boyle ,  Jose R. Brunheroto ,  Dong Chen ,  Chen-Yong Cher ,  George L. Chiu ,  Norman Christ ,  Paul W. Coteus ,  Kristan D. Davis ,  Gabor J. Dozsa ,  Alexandre E. Eichenberger ,  Noel A. Eisley ,  Matthew R. Ellavsky ,  Kahn C. Evans ,  Bruce M. Fleischer ,  Thomas W. Fox ,  Alan Gara ,  Mark E. Giampapa ,  Thomas M. Gooding ,  Michael K. Gschwind ,  John A. Gunnels ,  Shawn A. Hall ,  Rudolf A. Haring ,  Philip Heidelberger ,  Todd A. Inglett ,  Brant L. Knudson ,  Gerard V. Kopcsay ,  Sameer Kumar ,  Amith R. Mamidala ,  James A. Marcella ,  Mark G. Megerian ,  Douglas R. Miller ,  Samuel J. Miller ,  Adam J. Muff ,  Michael B. Mundy ,  John K. O'Brien ,  Kathryn M. O'Brien ,  Martin Ohmacht ,  Jeffrey J. Parker ,  Ruth J. Poole ,  Joseph D. Ratterman ,  Valentina Salapura ,  David L. Satterfield ,  Robert M. Senger ,  Burkhard Steinmacher-Burow ,  William M. Stockdell ,  Craig B. Stunkel ,  Krishnan Sugavanam ,  Yutaka Sugawara ,  Todd E. Takken ,  Barry M. Trager ,  James L. Van Oosten ,  Charles D. Wait ,  Robert E. Walkup ,  Alfred T. Watson ,  Robert W. Wisniewski ,  Peng Wu

IPC: G06F13/28 ,  G06F9/38 ,  G06F9/30 ,  G06F15/80 ,  G06F12/08

CPC classification number: G06F13/287 ,  G06F9/06 ,  G06F9/3004 ,  G06F9/30047 ,  G06F9/3885 ,  G06F12/0811 ,  G06F12/0831 ,  G06F12/0862 ,  G06F12/0864 ,  G06F12/1027 ,  G06F15/17381 ,  G06F15/17387 ,  G06F15/76 ,  G06F15/8069 ,  G06F2212/1016 ,  G06F2212/602 ,  G06F2212/6022 ,  G06F2212/6024 ,  G06F2212/6032 ,  Y02D10/13 ,  Y02D10/14

Abstract: A Multi-Petascale Highly Efficient Parallel Supercomputer of 100 petaflop-scale includes node architectures based upon System-On-a-Chip technology, where each processing node comprises a single Application Specific Integrated Circuit (ASIC). The ASIC nodes are interconnected by a five dimensional torus network that optimally maximize the throughput of packet communications between nodes and minimize latency. The network implements collective network and a global asynchronous network that provides global barrier and notification functions. Integrated in the node design include a list-based prefetcher. The memory system implements transaction memory, thread level speculation, and multiversioning cache that improves soft error rate at the same time and supports DMA functionality allowing for parallel processing message-passing.

Abstract translation: 100 petaflop规模的多千兆高效并行超级计算机包括基于片上系统技术的节点架构，其中每个处理节点包括单个专用集成电路（ASIC）。 ASIC节点通过五维环面网络互连，最优化节点之间的分组通信的吞吐量并最小化等待时间。 网络实现集体网络和提供全局障碍和通知功能的全球异步网络。 集成在节点设计中包括一个基于列表的预取器。 存储系统实现事务存储器，线程级别推测和多重切换缓存，同时提高软错误率，并支持DMA功能，允许并行处理消息传递。

公开(公告)号：US20130176904A1

公开(公告)日：2013-07-11

申请号：US13778581

申请日：2013-02-27

Applicant: INTERNATIONAL BUSINESS MACHINES CORPORATION

Inventor： Charles J. Archer ,  Daniel A. Faraj ,  Todd A. Inglett ,  Joseph D. Ratterman

IPC: H04L12/56

CPC classification number: H04L45/02 ,  G06F15/173

Abstract: Methods, apparatus, and products are disclosed for providing full point-to-point communications among compute nodes of an operational group in a global combining network of a parallel computer, each compute node connected to each adjacent compute node in the global combining network through a link, that include: receiving a network packet in a compute node, the network packet specifying a destination compute node; selecting, in dependence upon the destination compute node, at least one of the links for the compute node along which to forward the network packet toward the destination compute node; and forwarding the network packet along the selected link to the adjacent compute node connected to the compute node through the selected link.

Abstract translation: 公开了用于在并行计算机的全局组合网络中的操作组的计算节点之间提供完整点对点通信的方法，装置和产品，每个计算节点通过以下方式连接到全局组合网络中的每个相邻计算节点 链路，其包括：在计算节点中接收网络分组，所述网络分组指定目的地计算节点; 根据所述目的地计算节点选择用于所述计算节点的至少一个所述链路，用于将所述网络分组转发到所述目的地计算节点; 以及沿着所选择的链路将所述网络分组转发到通过所选择的链路连接到所述计算节点的相邻计算节点。

公开(公告)号：US20140223149A1

公开(公告)日：2014-08-07

申请号：US13785515

申请日：2013-03-05

Applicant: INTERNATIONAL BUSINESS MACHINES CORPORATION

Inventor： Todd A. Inglett ,  Andrew T. Tauferner

IPC: G06F9/30

CPC classification number: G06F7/58 ,  H04L9/0861 ,  H04L2209/125

Abstract: Method, system, and computer program product for randomizing entropy on a parallel computing system using network arithmetic logic units (ALUs). In one embodiment, network ALUs on nodes of the parallel computing system pseudorandomly modify entropy data during broadcast operations through application of arithmetic and/or logic operations. That is, each compute node's ALU may modify the entropy data during broadcasts, thereby mixing, and thus improving, the entropy data with every hop of entropy data packets from one node to another. At each compute node, the respective ALUs may further deposit modified entropy data in, e.g., local entropy pools such that software running on the compute nodes and needing entropy data may fetch it from the entropy pools. In some embodiments, entropy data may be broadcast via dedicated packets or included in unused portions of existing broadcast packets.

公开(公告)号：US09882801B2

公开(公告)日：2018-01-30

申请号：US13778581

申请日：2013-02-27

Applicant: INTERNATIONAL BUSINESS MACHINES CORPORATION

Inventor： Charles J. Archer ,  Daniel A. Faraj ,  Todd A. Inglett ,  Joseph D. Ratterman

IPC: H04L12/751 ,  G06F15/173

CPC classification number: H04L45/02 ,  G06F15/173

Abstract: Methods, apparatus, and products are disclosed for providing full point-to-point communications among compute nodes of an operational group in a global combining network of a parallel computer, each compute node connected to each adjacent compute node in the global combining network through a link, that include: receiving a network packet in a compute node, the network packet specifying a destination compute node; selecting, in dependence upon the destination compute node, at least one of the links for the compute node along which to forward the network packet toward the destination compute node; and forwarding the network packet along the selected link to the adjacent compute node connected to the compute node through the selected link.

公开(公告)号：US09335970B2

公开(公告)日：2016-05-10

申请号：US13785515

申请日：2013-03-05

Applicant: International Business Machines Corporation

Inventor： Todd A. Inglett ,  Andrew T. Tauferner

IPC: G06F7/58 ,  H04L9/08

CPC classification number: G06F7/58 ,  H04L9/0861 ,  H04L2209/125

Abstract: Method, system, and computer program product for randomizing entropy on a parallel computing system using network arithmetic logic units (ALUs). In one embodiment, network ALUs on nodes of the parallel computing system pseudorandomly modify entropy data during broadcast operations through application of arithmetic and/or logic operations. That is, each compute node's ALU may modify the entropy data during broadcasts, thereby mixing, and thus improving, the entropy data with every hop of entropy data packets from one node to another. At each compute node, the respective ALUs may further deposit modified entropy data in, e.g., local entropy pools such that software running on the compute nodes and needing entropy data may fetch it from the entropy pools. In some embodiments, entropy data may be broadcast via dedicated packets or included in unused portions of existing broadcast packets.

公开(公告)号：US09092285B2

公开(公告)日：2015-07-28

申请号：US13778715

申请日：2013-02-27

Applicant: International Business Machines Corporation

Inventor： Todd A. Inglett ,  Andrew T. Tauferner

IPC: G06F7/58

CPC classification number: G06F7/582 ,  G06F7/588

Abstract: Method for performing an operation, the operation including, responsive to receiving a file system request at a file system, retrieving a first entropy pool element from the file system, and inserting, at the file system, the first entropy pool element into a network packet sent from the file system responsive to the file system request.

Abstract translation: 用于执行操作的方法，所述操作包括响应于在文件系统处接收到文件系统请求，从所述文件系统检索第一熵池元素，以及在所述文件系统将所述第一熵池元素插入到网络包中 根据文件系统请求从文件系统发送。

公开(公告)号：US09569399B2

公开(公告)日：2017-02-14

申请号：US13668503

申请日：2012-11-05

Applicant: INTERNATIONAL BUSINESS MACHINES CORPORATION

Inventor： Charles J. Archer ,  Michael A. Blocksome ,  Todd A. Inglett ,  Joseph D. Ratterman ,  Brian E. Smith

IPC: H04L12/781 ,  H04L29/08 ,  G06F9/455 ,  G06F15/173

CPC classification number: G06F15/17312 ,  G06F9/45558 ,  H04L45/52 ,  H04L67/327

Abstract: Routing data communications packets in a parallel computer that includes compute nodes organized for collective operations. Each compute node including an operating system kernel and a system-level messaging module that is a module of automated computing machinery that exposes a messaging interface to applications. Each compute node including a routing table that specifies, for each of a multiplicity of route identifiers, a data communications path through the compute node. Including to carry out the steps of: receiving in a compute node a data communications packet that includes a route identifier value; retrieving from the routing table a specification of a data communications path through the compute node; and routing, by the compute node, the data communications packet according to the data communications path identified by the compute node's routing table entry for the data communications packet's route identifier value.

Abstract translation: 在包括为集体操作组织的计算节点的并行计算机中路由数据通信数据包。 每个计算节点包括操作系统内核和系统级消息传递模块，该模块是将应用程序的消息传递接口公开的自动化计算机的模块。 每个计算节点包括路由表，其为多个路由标识符中的每一个指定通过计算节点的数据通信路径。 包括以下步骤：在计算节点中接收包括路由标识符值的数据通信分组; 从所述路由表中检索通过所述计算节点的数据通信路径的规范; 并且由计算节点根据数据通信分组的路由标识符值的计算节点的路由表项标识的数据通信路径路由数据通信分组。

公开(公告)号：US09335969B2

公开(公告)日：2016-05-10

申请号：US13760510

申请日：2013-02-06

Applicant: International Business Machines Corporation

Inventor： Todd A. Inglett ,  Andrew T. Tauferner

IPC: G06F7/58 ,  H04L9/08

CPC classification number: G06F7/58 ,  H04L9/0861 ,  H04L2209/125

Abstract: Method, system, and computer program product for randomizing entropy on a parallel computing system using network arithmetic logic units (ALUs). In one embodiment, network ALUs on nodes of the parallel computing system pseudorandomly modify entropy data during broadcast operations through application of arithmetic and/or logic operations. That is, each compute node's ALU may modify the entropy data during broadcasts, thereby mixing, and thus improving, the entropy data with every hop of entropy data packets from one node to another. At each compute node, the respective ALUs may further deposit modified entropy data in, e.g., local entropy pools such that software running on the compute nodes and needing entropy data may fetch it from the entropy pools. In some embodiments, entropy data may be broadcast via dedicated packets or included in unused portions of existing broadcast packets.

Abstract translation: 使用网络算术逻辑单元（ALU）在并行计算系统上随机化熵的方法，系统和计算机程序产品。 在一个实施例中，并行计算系统的节点上的网络ALU在广播操作期间通过应用算术和/或逻辑运算来伪随机地修改熵数据。 也就是说，每个计算节点的ALU可以在广播期间修改熵数据，从而将熵数据与来自一个节点的熵数据分组的每一跳混合并从而改善熵数据。 在每个计算节点处，相应的ALU可以进一步在例如本地熵池中存储修改的熵数据，使得在计算节点上运行并需要熵数据的软件可以从熵池获取它。 在一些实施例中，熵数据可以经由专用分组广播或包括在现有广播分组的未使用部分中。

公开(公告)号：US08745114B2

公开(公告)日：2014-06-03

申请号：US13684073

申请日：2012-11-21

Applicant: International Business Machines Corporation

Inventor： Todd A. Inglett ,  Andrew T. Tauferner

IPC: G06F7/58

CPC classification number: G06F7/582 ,  G06F7/588

Abstract: A parallel computer system adds entropy to improve the quality of random number generation by using parity errors as a source of entropy because parity errors are influenced by external forces such as cosmic ray bombardment, alpha particle emission, and other random or near-random events. By using parity errors and associated information to generate entropy, the quality of random number generation in a parallel computer system is increased.

Abstract translation: 并行计算机系统通过使用奇偶校验错误作为熵源来增加熵来提高随机数生成的质量，因为奇偶校验错误受外力如宇宙射线轰击，α粒子发射和其他随机或近随机事件的影响。 通过使用奇偶校验错误和相关信息来产生熵，并行计算机系统中随机数生成的质量得到提高。

公开(公告)号：US20140040336A1

公开(公告)日：2014-02-06

申请号：US13778715

申请日：2013-02-27

Applicant: INTERNATIONAL BUSINESS MACHINES CORPORATION

Inventor： Todd A. Inglett ,  Andrew T. Tauferner

IPC: G06F7/58

CPC classification number: G06F7/582 ,  G06F7/588

Abstract: Method for performing an operation, the operation including, responsive to receiving a file system request at a file system, retrieving a first entropy pool element from the file system, and inserting, at the file system, the first entropy pool element into a network packet sent from the file system responsive to the file system request.

Abstract translation: 用于执行操作的方法，所述操作包括响应于在文件系统处接收到文件系统请求，从所述文件系统检索第一熵池元素，以及在所述文件系统将所述第一熵池元素插入到网络包中 根据文件系统请求从文件系统发送。