Abstract: A method for diagnosing communications includes sending a message from a sending node to a receiving node. The sending node detects an error in the receiving node receiving the message. A force log request is sent from the sending node to the receiving node, the force log request including a request for the receiving node to log information. A force log response is received from the receiving node at the sending node, the force log response including the logged information. The sending node diagnoses the communications error in response to the force log response.

Abstract: A method of providing a converged InfiniBand over Ethernet (IBOE) network. An IBOE switch receives a data packet. If the data packet is received by the IBOE switch from the InfiniBand network, a translation utility looks up an Ethernet media access control (MAC) address corresponding to an LID of the packet in a translation table. The translation utility generates an Ethernet packet and encapsulates an InfiniBand link layer packet before sending the packet. If the data packet is received by the IBOE switch from the Ethernet network, the translation utility removes the Ethernet header from the packet and looks up a LID corresponding to the MAC address. The translation utility calculates an outbound port number and sends the packet.

Abstract: A method of providing visibility of Ethernet components to a subnet manager in a converged InfiniBand over Ethernet (IBOE) network. If a port of an IBOE gateway corresponds to one or more InfiniBand devices, the subnet manager sends fabric management packets (FMPs) to discover the InfiniBand network and assigns physical local identifiers (LIDs) to the InfiniBand devices. If a port of the IBOE gateway corresponds to one or more Ethernet devices, the subnet manager sends FMPs to discover the Ethernet network. The subnet manager adds the Ethernet Media Access Control (MAC) addresses of any responding devices to an LID routing table and assigns LIDs to the Ethernet devices. The subnet manager configures one or more virtual Host Channel Adapters (HCAs) corresponding to the one or more Ethernet MAC addresses in the LID routing table.

Abstract: A method of providing a converged InfiniBand over Ethernet (IBOE) network. An IBOE switch receives a data packet. If the data packet is received by the IBOE switch from the InfiniBand network, a translation utility generates an Ethernet media access control (MAC) address corresponding to an InfiniBand Local Identifier (LID) of said data packet. The translation utility generates an Ethernet packet and encapsulates an InfiniBand link layer packet before sending the packet. If the data packet is received by the IBOE switch from the Ethernet network, the translation utility removes the Ethernet header from the packet and looks up a LID corresponding to the MAC address. The translation utility calculates an outbound port number and sends the packet.

Abstract: A system and method for providing a memory region/memory window (MR/MW) access notification on a system area network are provided. Whenever a previously allocated MR/MW is accessed, such as via a remote direct memory access (RDMA) read/write operation, a notification of the access is generated and written to a queue data structure associated with the MR/MW. In one illustrative embodiment, this queue data structure may be a MR/MW event queue (EQ) data structure that is created and used for all consumer processes and all MR/MWs. In other illustrative embodiments, the EQ is associated with a protection domain. In yet another illustrative embodiment, an event record may be posted to an asynchronous event handler in response to the accessing of the MR/MW. In another illustrative embodiment, a previously posted queue element may be used to generate a completion queue element in response to the accessing of the MR/MW.

Abstract: A method for supervisor partitioning of client resources in a communications environment includes receiving, at a client node, an allocated set of resources over a communications network, and partitioning the allocated set of resources among one or more applications associated with the client node using a local authority. Following the partitioning, communication requests are issued from the one or more applications to a shared resource provider node without inspection by the local authority.

Abstract: Resources of a server node are logically divided into a plurality of sets of resources. At least one set of resources is assigned to one or more client nodes. The association of the at least one set of resources with the one or more client nodes is via a data structure stored at the server node. The data structure is provided by a trusted agent over a communications medium coupling the server node and the one or more client nodes. It includes information that indicates the resources accessible by the client nodes. To access the information, an identifier, also provided by a trusted agent, is employed.

Abstract: A method for authenticating a requesting entity in a communications environment. In an exemplary embodiment, the method includes determining a client identification of a client node associated with the requesting entity, and determining whether the requesting entity associated with the client node is acting in a supervisor capacity. A key to the requesting entity is returned from a resource provider node upon determining that the client identification of the client node indicates that the client node is permitted to access one or more resources of the provider node, and that the client node is acting in a supervisor capacity.

Abstract: A system for authenticating a requesting entity in a subnet communications environment includes determining a client identification of a client node associated with the requesting entity, and determining whether the requesting entity associated with the client node is acting in a supervisor capacity. A key to the requesting entity is returned from a resource provider node upon determining that the client identification of the client node indicates that the client node is permitted to access one or more resources of the provider node, and that the client node is acting in a supervisor capacity.

Abstract: A method for authenticating a requesting entity in a communications environment. In an exemplary embodiment, the method includes determining a client identification of a client node associated with the requesting entity, and determining whether the requesting entity associated with the client node is acting in a supervisor capacity. A key to the requesting entity is returned from a resource provider node upon determining that the client identification of the client node indicates that the client node is permitted to access one or more resources of the provider node, and that the client node is acting in a supervisor capacity.

Abstract: A system for supervisor partitioning of client resources in a subnet communications environment includes a plurality of client nodes, each receiving an allocated set of resources determined by a central authority, which assigns resources to a supervisor key associated with each supervisor of the plurality of client nodes. At each of the client nodes, the allocated set of resources are partitioned using a local supervisor associated therewith, with each supervisor associating one or more resource keys with one or more resources allocated to the corresponding client node, and assigning the one or more resource keys to the one or more applications. The supervisor keys are configured so as to prevent a given supervisor from partitioning resources not allocated to the client node associated therewith. Following the partitioning, communication packets are issued from the one or more applications to a resource provider node without inspection by the corresponding supervisor.

Abstract: A system and method for collective send operations on a system area network are provided. The mechanisms of the illustrative embodiments provide for the creation, modification, and removal of collective send queues (CSQs) that allow the upper level protocol (ULP) used by a consumer to send the same message to a collective set of queue pairs (QPs). In order to use the transport services of a CSQ, a consumer process posts a write work request (WR) to the CSQ. The write WR causes a write work queue element (WQE) to be generated and placed in the CSQ. A channel interface (CI) is provided that effectively copies the write WQE to all of the send queues (SQs) of the QPs in the QP set associated with the CSQ. When all the QPs complete processing of their respective write WQEs, the HCA releases all data segments referenced by the write WR.

Abstract: A distributed computer system includes a first subnet including a first group of endnodes, each endnode in the first group of endnodes having a unique destination location identification (DLID) within the first subnet. A second subnet in the distributed computer system includes a second group of endnodes, each endnode in the second group of endnodes having a unique DLID within the second subnet. A communication fabric is physically coupled to the first group of endnodes and the second group of endnodes. A partitioning mechanism associates a first partition key to every DLID in the first subnet for enabling communication between the first group of endnodes over the communication fabric. The partitioning mechanism associates a second partition key to every DLID in the second subnet for enabling communication between the second group of endnodes over the communication fabric.

Abstract: An electronic computer system including a central processor and a hierarchical memory system having a large relatively low speed random access system memory and a small high speed set-associative cache memory including a data store section for storing lines of data from the system memory and a cache directory for indicating, by means of line identifier fields at any time, the lines of the system memory data currently resident in cache, is provided with a way to improve the distribution of data across the congruence classes within the cache. A mechanism is provided for performing a permutation operation on an M bit portion (X) of the system memory address, which permutation determines the congruence class into which the address will map. The permutation mechanism performs a bit-matrix multiplication of said M-bit address with an M.times.M matrix (where M is a real positive integer greater than 1) to produce a permuted M-bit address (X').

Abstract: A multiprocessing system is presented having a plurality of processing nodes interconnected together by a communication network, each processing node including a processor, responsive to user software running on the system, and an associated memory module, and capable under user control of dynamically partitioning each memory module into a global storage efficiently accessible by a number of processors connected to the network, and local storage efficiently accessible by its associated processor.

Abstract: An interconnection network management architecture for use with a large shared memory multiprocessor computing system including a plurality of processors and a plurality of separately addressable main memory modules. Two parallel, interconnection networks are provided each capable of interconnecting any processor to any memory module, and each having different latency characteristics. A Hot-Spot detection mechnaism is associated with each main memory module for detecting when a particular address in that module has become a Hot Spot and includes a first memory for storing all detected Hot Spots. A diverter element is associated with each processor for selectively routing memory requests over either the first or second memory network contingent on its status as a Hot Spot. A second memory is included in each diverter element for storing all Hot Spots detected by the detector elements.

Abstract: A method and apparatus for marking data that is temporarily cacheable to facilitate the efficient management of said data. A bit in the segment and/or page descriptor of the data called the marked data bit (MDB) is generated by the compiler and included in a request for data from memory by the processor in the form of a memory address and will be stored in the cache directory at a location related to the particular line of data involved. The bit is passed to the cache together with the associated real address after address translation (in the case of a real cache). when the cache controls load the address of the data in the directory it is also stored the marked data bit (MDB) in the directory with the address. When the cacheability of the temporarily cacheable data changes from cacheable to non-cacheable, a single instruction is issued to cause the cache to invalidate all marked data.

Abstract: A multiprocessing system is presented for dynamically partitioning a storage module into a global storage efficiently accessible by a number of processors connected to a network, and local storage efficiently accessible by individual processors, including the interleaving of storage references output by a processor, under the control of that processor, and dynamically directing the storage references to first or second portions of the storage module.