公开(公告)号：US20140052954A1

公开(公告)日：2014-02-20

申请号：US13969425

申请日：2013-08-16

申请人： Laurent MOLL ,  Jean-Jacques LECLER ,  Philippe BOUCARD

发明人： Laurent MOLL ,  Jean-Jacques LECLER ,  Philippe BOUCARD

IPC分类号： G06F12/02 ,  G06F12/10

CPC分类号： G06F12/1036 ,  G06F12/0207 ,  G06F12/0215 ,  G06F12/0811 ,  G06F12/1027 ,  G06F12/1081 ,  G06F17/5009 ,  G06F2212/681

摘要： A system TLB accepts translation prefetch requests from initiators. Misses generate external translation requests to a walker port. Attributes of the request such as ID, address, and class, as well as the state of the TLB affect the allocation policy of translations within multiple levels of translation tables. Translation tables are implemented with SRAM, and organized in groups.

摘要翻译： 系统TLB接受来自发起者的翻译预取请求。 Misses会向Walker端口生成外部翻译请求。 诸如ID，地址和类的请求的属性以及TLB的状态会影响多级翻译表中的翻译分配策略。 翻译表用SRAM实现，并组织成组。

公开(公告)号：US20100296400A1

公开(公告)日：2010-11-25

申请号：US12848938

申请日：2010-08-02

申请人： Cesar Douady ,  Philippe Boucard

发明人： Cesar Douady ,  Philippe Boucard

IPC分类号： H04L12/24

CPC分类号： H04L47/6215 ,  H04L47/19 ,  H04L47/24 ,  H04L47/50 ,  H04L47/60 ,  H04L2012/5651

摘要： Method of managing priority during the transmission of a message, in an interconnections network comprising at least one transmission agent which comprises at least one input and at least one output, each input comprising a means of storage organized as a queue of messages. A message priority is assigned during the creation of the message, and a queue priority equal to the maximum of the priorities of the messages of the queue is assigned to at least one queue of messages of an input. A link priority is assigned to a link linking an output of a first transmission agent to an input of a second transmission agent, equal to the maximum of the priorities of the queues of messages of the inputs of said first agent comprising a first message destined for that output of said first agent which is coupled to said link, and the priority of the link is transmitted to that input of said second agent which is coupled to the link.

摘要翻译： 在包括至少一个传输代理的互连网络中管理消息传输期间的优先级的方法，所述至少一个传输代理包括至少一个输入和至少一个输出，每个输入包括组织为消息队列的存储装置。 在创建消息期间分配消息优先级，并且将等于队列消息的优先级的最大值的队列优先级分配给至少一个输入的消息队列。 将链路优先级分配给链接第一传输代理的输出到第二传输代理的输入的链路，等于所述第一代理的输入的消息的队列的优先级的最大值，该第一代理的输入的消息队列的优先级最大值包括目的地为 耦合到所述链路的所述第一代理的输出，并且链路的优先级被发送到耦合到链路的所述第二代理的输入。

公开(公告)号：US20100122004A1

公开(公告)日：2010-05-13

申请号：US12616694

申请日：2009-11-11

申请人： Philippe Boucard ,  Luc Montperrus

发明人： Philippe Boucard ,  Luc Montperrus

IPC分类号： G06F13/38

CPC分类号： H04L47/58 ,  H04L49/109 ,  H04L49/254

摘要： The message switching system comprises at least two inputs and at least one output, first arbitration means dedicated to said output, and management means designed to determine a relative order OR(i,j) of one input relative to the other, for any pair of separate inputs belonging to the system and having sent requests for the assignment of said output, and designed to assign said output. Said management means comprise storage means designed to store said relative orders OR(i,j), initialization means designed to initialize said relative orders OR(i,j) such that only one of said inputs takes priority on initialization, and updating means designed to update all of said relative orders when a new request arrives at said first arbitration means, or when said output is assigned to one of said inputs.

摘要翻译： 消息交换系统包括至少两个输入和至少一个输出，专用于所述输出的第一仲裁装置和被设计为确定一个输入相对于另一个输入的相对顺序OR（i，j）的管理装置， 单独的输入属于系统并且已经发送了用于分配所述输出的请求，并被设计成分配所述输出。 所述管理装置包括设计用于存储所述相对命令OR（i，j）的存储装置，被设计为初始化所述相对命令OR（i，j）的初始化装置，使得只有一个所述输入在初始化时优先，并且更新装置被设计为 当新请求到达所述第一仲裁装置时，或者当所述输出被分配给所述输入之一时，更新所有所述相对命令。

公开(公告)号：US07639704B2

公开(公告)日：2009-12-29

申请号：US11484837

申请日：2006-07-11

申请人： Philippe Boucard ,  Luc Montperrus

发明人： Philippe Boucard ,  Luc Montperrus

IPC分类号： H04L12/28 ,  H04L12/56

CPC分类号： H04L47/58 ,  H04L49/109 ,  H04L49/254

摘要： The message switching system (51) comprises at least two inputs (52, 53, 54, 55) and at least one output (56), first arbitration means (62) dedicated to said output (56), and management means (64) designed to determine a relative order OR(i,j) of one input relative to the other, for any pair of separate inputs belonging to the system (51) and having sent requests for the assignment of said output (56), and designed to assign said output (56). Said management means (64) comprise storage means (70) designed to store said relative orders OR(i,j), initialization means (66) designed to initialize said relative orders OR(i,j) such that only one of said inputs takes priority on initialization, and updating means (68) designed to update all of said relative orders when a new request arrives at said first arbitration means (62), or when said output is assigned to one of said inputs.

摘要翻译： 消息交换系统（51）包括至少两个输入（52,53,54,55）和至少一个输出（56），专用于所述输出（56）的第一仲裁装置（62）和管理装置（64） 被设计为为属于系统（51）的任何一对单独输入确定一个输入相对于另一输入的相对顺序OR（i，j），并且已经发送了对所述输出（56）的分配的请求，并被设计为 分配所述输出（56）。 所述管理装置（64）包括设计成存储所述相对命令OR（i，j）的存储装置（70），被设计为初始化所述相对命令OR（i，j）的初始化装置（66），使得只有一个所述输入 以及当新的请求到达所述第一仲裁装置（62）时或者当所述输出被分配给所述输入之一时被设计为更新所有所述相对命令的更新装置（68）。

公开(公告)号：US20050210325A1

公开(公告)日：2005-09-22

申请号：US11054179

申请日：2005-02-09

申请人： Cesar Douady ,  Philippe Boucard

发明人： Cesar Douady ,  Philippe Boucard

IPC分类号： G06F11/00 ,  G06F13/38 ,  G06F13/40 ,  H04L12/56 ,  H04L12/58

CPC分类号： G06F13/4027

摘要： In some embodiments, a system for communication between agents in a point-to-point interconnection system includes at least one initiator agent capable of dispatching at least one message destined for at least one determined receiver agent; at least one intermediate agent capable of forwarding at least one message to at least one determined receiver agent; at least one receiver agent capable of receiving at least one message originating from an initiator agent via at least one intermediate agent; and means of error detection and means of erroneous message marking. In some embodiments, a receiver agent includes means for formulating an error message and means for sending the error message to the initiator agent so as to warn the said initiator agent of the presence of an error.

摘要翻译： 在一些实施例中，用于在点对点互连系统中的代理之间的通信的系统包括能够调度至少一个目的地为至少一个确定的接收方代理的消息的至少一个发起方代理; 至少一个能够将至少一个消息转发到至少一个确定的接收方代理的中间代理; 至少一个能够通过至少一个中间代理接收来自起始代理的至少一个消息的接收方代理; 以及错误检测的手段和错误信息标记的手段。 在一些实施例中，接收方代理包括用于制定错误消息的装置和用于向发起方代理发送错误消息的装置，以便警告所述发起方代理存在错误。

公开(公告)号：US20130166812A1

公开(公告)日：2013-06-27

申请号：US13337280

申请日：2011-12-26

申请人： PHILIPPE BOUCARD ,  JEAN-JACQUES LECLER ,  PHILIPPE MARTIN ,  LAURENT MOLL

发明人： PHILIPPE BOUCARD ,  JEAN-JACQUES LECLER ,  PHILIPPE MARTIN ,  LAURENT MOLL

IPC分类号： G06F13/36

CPC分类号： G06F13/4022 ,  G06F13/405

摘要： A system and method are disclosed for connecting PCI-ordered agents based on fully independent networks. The system and method are free of PCI topology constraints, so that the system and method can be implemented in an inexpensive and scalable way. The method disclosed is used to handle and transport PCI-ordered traffic on a fabric. Based on the actual ordering requirement of the set of PCI agents, the fabric includes two, three, or four independent networks.

摘要翻译： 公开了一种用于连接基于完全独立网络的PCI有序代理的系统和方法。 该系统和方法没有PCI拓扑约束，从而可以以便宜且可扩展的方式实现系统和方法。 所公开的方法用于在结构上处理和传输PCI-有序流量。 基于一组PCI代理的实际排序要求，该架构包括两个，三个或四个独立网络。

公开(公告)号：US20110302345A1

公开(公告)日：2011-12-08

申请号：US12835623

申请日：2010-07-13

申请人： Philippe Boucard ,  Philippe Martin ,  Jean-Jacques Lecler

发明人： Philippe Boucard ,  Philippe Martin ,  Jean-Jacques Lecler

IPC分类号： G06F13/368

CPC分类号： G06F13/362 ,  H04L45/302 ,  H04L47/10 ,  H04L47/2441 ,  H04L47/2458 ,  H04L49/10 ,  H04L49/109 ,  H04L49/205

摘要： Quality-of-Service (QoS) is an important system-level requirement in the design and implementation of on-chip networks. QoS requirements can be implemented in an on-chip-interconnect by providing for at least two signals indicating priority at a transaction-level interface where one signal transfers information in-band with the transaction and the other signal transfers information out-of-band with the transaction. The signals can be processed by the on-chip-interconnect to deliver the required QoS. In addition, the disclosed embodiments can be extended to a Network-on-Chip (NoC).

摘要翻译： 服务质量（QoS）是片上网络设计和实现中的重要系统级要求。 QoS要求可以通过提供至少两个表示交易级别接口处的优先级的信号来实现，其中一个信号在事务处理带内传输信息，另一个信号在带外传输信息， 交易。 这些信号可以通过片上互连来处理以提供所需的QoS。 另外，所公开的实施例可以扩展到片上网络（NoC）。

公开(公告)号：US07769027B2

公开(公告)日：2010-08-03

申请号：US10915950

申请日：2004-08-11

申请人： Cesar Douady ,  Philippe Boucard

发明人： Cesar Douady ,  Philippe Boucard

IPC分类号： H04L12/56

CPC分类号： H04L47/6215 ,  H04L47/19 ,  H04L47/24 ,  H04L47/50 ,  H04L47/60 ,  H04L2012/5651

摘要： Method of managing priority during the transmission of a message, in an interconnections network comprising at least one transmission agent which comprises at least one input and at least one output, each input comprising a means of storage organized as a queue of messages. A message priority is assigned during the creation of the message, and a queue priority equal to the maximum of the priorities of the messages of the queue is assigned to at least one queue of messages of an input. A link priority is assigned to a link linking an output of a first transmission agent to an input of a second transmission agent, equal to the maximum of the priorities of the queues of messages of the inputs of said first agent comprising a first message destined for that output of said first agent which is coupled to said link, and the priority of the link is transmitted to that input of said second agent which is coupled to the link.

摘要翻译： 在包括至少一个传输代理的互连网络中管理消息传输期间的优先级的方法，所述至少一个传输代理包括至少一个输入和至少一个输出，每个输入包括组织为消息队列的存储装置。 在创建消息期间分配消息优先级，并且将等于队列消息的优先级的最大值的队列优先级分配给至少一个输入的消息队列。 将链路优先级分配给链接第一传输代理的输出到第二传输代理的输入的链路，等于所述第一代理的输入的消息的队列的优先级的最大值，该第一代理的输入的消息队列的优先级最大值包括目的地为 耦合到所述链路的所述第一代理的输出，并且链路的优先级被发送到耦合到链路的所述第二代理的输入。

公开(公告)号：US20050117589A1

公开(公告)日：2005-06-02

申请号：US10915950

申请日：2004-08-11

申请人： Cesar Douady ,  Philippe Boucard

发明人： Cesar Douady ,  Philippe Boucard

IPC分类号： H04L12/28 ,  H04L12/46 ,  H04L12/54 ,  H04L12/58 ,  H04L29/02

CPC分类号： H04L47/6215 ,  H04L47/19 ,  H04L47/24 ,  H04L47/50 ,  H04L47/60 ,  H04L2012/5651

摘要： Method of managing priority during the transmission of a message, in an interconnections network comprising at least one transmission agent which comprises at least one input and at least one output, each input comprising a means of storage organized as a queue of messages. A message priority is assigned during the creation of the message, and a queue priority equal to the maximum of the priorities of the messages of the queue is assigned to at least one queue of messages of an input. A link priority is assigned to a link linking an output of a first transmission agent to an input of a second transmission agent, equal to the maximum of the priorities of the queues of messages of the inputs of said first agent comprising a first message destined for that output of said first agent which is coupled to said link, and the priority of the link is transmitted to that input of said second agent which is coupled to the link.

摘要翻译： 在包括至少一个传输代理的互连网络中管理消息传输期间的优先级的方法，所述至少一个传输代理包括至少一个输入和至少一个输出，每个输入包括组织为消息队列的存储装置。 在创建消息期间分配消息优先级，并且将等于队列消息的优先级的最大值的队列优先级分配给至少一个输入的消息队列。 将链路优先级分配给链接第一传输代理的输出到第二传输代理的输入的链路，等于所述第一代理的输入的消息的队列的优先级的最大值，该第一代理的输入的消息队列的优先级最大值包括目的地为 耦合到所述链路的所述第一代理的输出，并且链路的优先级被发送到耦合到链路的所述第二代理的输入。

公开(公告)号：US09069912B2

公开(公告)日：2015-06-30

申请号：US13436944

申请日：2012-03-31

申请人： Philippe Boucard ,  Jean-Jacques Lecler

发明人： Philippe Boucard ,  Jean-Jacques Lecler

IPC分类号： G06F13/36 ,  G06F13/40 ,  G06F13/22

CPC分类号： G06F13/4059 ,  G06F13/22 ,  G06F13/4031

摘要： A Network-on-Chip (NoC) is provided that performs reordering of transaction responses such as those with requests that cross address mapping boundaries. Ordering is ensured by filtering logic in reorder buffers, some of which include storage to allow multiple simultaneously pending transactions. Transactions are transported by a packet-based transport protocol. The reorder buffering is done at packet level, within the transport topology. Reorder buffers are distributed physically throughout the floorplan of the chip, they have localized connectivity to initiators, and they operate in separate power and clock domains.

摘要翻译： 提供了片上网络（NoC），其执行诸如具有跨地址映射边界的请求的事务响应的重新排序。 通过在重排序缓冲区中过滤逻辑来确保排序，其中一些包括存储以允许多个同时挂起的事务。 事务通过基于分组的传输协议传送。 重新排序缓冲在传输拓扑中的分组级完成。 重新排序缓冲区物理地分布在芯片的整个平面图中，它们具有到启动器的本地化连接，并且它们在单独的电源和时钟域中操作。