Abstract: Providing memory bandwidth compression using multiple last-level cache (LLC) lines in a central processing unit (CPU)-based system is disclosed. In some aspects, a compressed memory controller (CMC) provides an LLC comprising multiple LLC lines, each providing a plurality of sub-lines the same size as a system cache line. The contents of the system cache line(s) stored within a single LLC line are compressed and stored in system memory within the memory sub-line region corresponding to the LLC line. A master table stores information indicating how the compressed data for an LLC line is stored in system memory by storing an offset value and a length value for each sub-line within each LLC line. By compressing multiple system cache lines together and storing compressed data in a space normally allocated to multiple uncompressed system lines, the CMC enables compression sizes to be smaller than the memory read/write granularity of the system memory.

Abstract: Providing memory bandwidth compression using multiple last-level cache (LLC) lines in a central processing unit (CPU)-based system is disclosed. In some aspects, a compressed memory controller (CMC) provides an LLC comprising multiple LLC lines, each providing a plurality of sub-lines the same size as a system cache line. The contents of the system cache line(s) stored within a single LLC line are compressed and stored in system memory within the memory sub-line region corresponding to the LLC line. A master table stores information indicating how the compressed data for an LLC line is stored in system memory by storing an offset value and a length value for each sub-line within each LLC line. By compressing multiple system cache lines together and storing compressed data in a space normally allocated to multiple uncompressed system lines, the CMC enables compression sizes to be smaller than the memory read/write granularity of the system memory.

Abstract: Providing memory bandwidth compression using multiple last-level cache (LLC) lines in a central processing unit (CPU)-based system is disclosed. In some aspects, a compressed memory controller (CMC) provides an LLC comprising multiple LLC lines, each providing a plurality of sub-lines the same size as a system cache line. The contents of the system cache line(s) stored within a single LLC line are compressed and stored in system memory within the memory sub-line region corresponding to the LLC line. A master table stores information indicating how the compressed data for an LLC line is stored in system memory by storing an offset value and a length value for each sub-line within each LLC line. By compressing multiple system cache lines together and storing compressed data in a space normally allocated to multiple uncompressed system lines, the CMC enables compression sizes to be smaller than the memory read/write granularity of the system memory.

Abstract: Providing memory bandwidth compression using multiple last-level cache (LLC) lines in a central processing unit (CPU)-based system is disclosed. In some aspects, a compressed memory controller (CMC) provides an LLC comprising multiple LLC lines, each providing a plurality of sub-lines the same size as a system cache line. The contents of the system cache line(s) stored within a single LLC line are compressed and stored in system memory within the memory sub-line region corresponding to the LLC line. A master table stores information indicating how the compressed data for an LLC line is stored in system memory by storing an offset value and a length value for each sub-line within each LLC line. By compressing multiple system cache lines together and storing compressed data in a space normally allocated to multiple uncompressed system lines, the CMC enables compression sizes to be smaller than the memory read/write granularity of the system memory.

Abstract: Providing memory bandwidth compression using multiple last-level cache (LLC) lines in a central processing unit (CPU)-based system is disclosed. In some aspects, a compressed memory controller (CMC) provides an LLC comprising multiple LLC lines, each providing a plurality of sub-lines the same size as a system cache line. The contents of the system cache line(s) stored within a single LLC line are compressed and stored in system memory within the memory sub-line region corresponding to the LLC line. A master table stores information indicating how the compressed data for an LLC line is stored in system memory by storing an offset value and a length value for each sub-line within each LLC line. By compressing multiple system cache lines together and storing compressed data in a space normally allocated to multiple uncompressed system lines, the CMC enables compression sizes to be smaller than the memory read/write granularity of the system memory.

Abstract: Computers are caused to provide a hash table wherein each entry is associated with a binary key and indexed by a selected portion of a hash value of the associated key, and points to a data structure location for storing non-selected portions of, or the entire hash value of, the binary key, and action data corresponding to the value of the binary key. Content addressable memory entries store a binary key, or a value unique to it, and an association to a corresponding action. Pointers to the data structure use selected portions of binary key hash values as an index when not selected portions of hash values of other binary keys, and associations are established between CAM entry and associated data structure locations when selected portions of the hash values of the binary keys are the same as selected portions of hash values of one or more other binary keys.

Abstract: A Resource Reservation System includes a Token Generation Unit (TGU) which generates and circulates among nodes of a communications system a Slotted Token (SLT) message having sub-fields to carry identification number for each input port in a node and the resource available for each input port. On receiving the message the Resource Control Unit (RCU) in each node can write port identification number, available resource in appropriate sub-fields of the SLT message, and reserve resources in other nodes by adjusting information in the sub-field associated with the other nodes.

Abstract: A structure and technique for preventing collisions using a hash table in conjunction with a CAM to identify and prevent collision of binary keys. A portion of the hash value of a binary key, which does not collide with a portion of the hash value of any other reference binary key, is used as an entry in the hash table. If two or more binary keys have identical values of the portions of the hash values, each of these binary keys are stored in their entirety, in the CAM. The key in the CAM provides a pointer to a data structure where the action associated with that binary key is stored. If the binary key is not found in the CAM, the binary key is hashed, and a specific entry in the hash table is selected using a portion of this hash value.

Abstract: A structure and technique for preventing collisions using a hash table in conjunction with a CAM to identify and prevent collisions of binary keys. A portion of the hash value of a binary key, which does not collide with a portion of the hash value of any other reference binary key, is used as an entry in the hash table. If two or more binary keys have identical values of the portions of the hash values, each of these binary keys are stored in their entirety, in the CAM. The key in the CAM provides a pointer to a data structure where the action associated with that binary key is stored. If the binary key is not found in the CAM, the binary key is hashed, and a specific entry in the hash table is selected using a portion of this hash value.

Abstract: A Resource Reservation System includes a Token Generation Unit (TGU) which generates and circulates among nodes of a communications system a Slotted Token (SLT) message having sub-fields to carry identification number for each input port in a node and the resource available for each input port. On receiving the message the Resource Control Unit (RCU) in each node can write port identification number, available resource in appropriate sub-fields of the SLT message, and reserve resources in other nodes by adjusting information in the sub-field associated with the other nodes.

Abstract: A structure and technique for preventing collisions using a hash table in conjunction with a CAM to identify and prevent collisions of binary keys. A portion of the hash value of a binary key, which does not collide with a portion of the hash value of any other reference binary key, is used as an entry in the hash table. If two or more binary keys have identical values of the portions of the hash values, each of these binary keys are stored in their entirety, in the CAM. The key in the CAM provides a pointer to a data structure where the action associated with that binary key is stored. If the binary key is not found in the CAM, the binary key is hashed, and a specific entry in the hash table is selected using a portion of this hash value.

Abstract: A method and apparatus useful in network management which makes intelligent, high speed, connection allocation decisions, overcoming difficulties encountered heretofore and providing enhanced network services. During episodes of network congestion, some connection requests for a class of service of low value and with currently a high number of existing connections may be purposefully ignored (not acknowledged with an Acknowledge (ACK) packet) so that the processing capability of a device will not become overwhelmed, causing the dropping of new connection is to note the numbers of connections of different classes relative to their service-level contracts, to ignore abundant, low-value connection requests in accordance with value policies when and only when necessary, and to insure that valuable new connection requests that conform to their contract connection rates can be intelligently accommodated.

Abstract: Novel data structures, methods and apparatus for a Software Managed Tree (SMT) which provides a mechanism to create tree structures that follow a search mechanism defined by a control point processor. The search mechanism does not require storage on the previous pointer and uses only a forward pointer along with a next bit or group of bits to test thereby reducing storage space for nodes. The search mechanism processes multiple filter rules for an application without requiring multiple searches and also allows various filter rules to be chained. Two patterns of the same length are stored in each leaf to define a range compare. A compare at the end operation is either a compare under range or a compare under mask. In a compare under range, the input key is checked to determine if it is in the range defined by the two patterns. In a compare under mask, the bits in the input key are compared with the bits in a first leaf pattern under a mask specified in a second leaf pattern.

Abstract: The transport protocol for communicating between general purpose processors acting as contact points and network processors in a packet processing environment such as Ethernet is provided. In such an environment, there is at least one single control point processor (CP) and a plurality of network processors (NP), sometimes referred to as blades. A typical system could contain two to sixteen network processors, and each network processor connects to a plurality of devices which communicate with each other over a network transport, such as Ethernet. The CP typically controls the functionality and the functioning of the network processors to function in a way that connects one end user with another, whether or not the end user is on the same network processor or a different network processor.

Abstract: The decision to discard or forward a packet is made by a flow control mechanism, upstream from the forwarding engine in the node of a communication network. The forwarding engine includes a switch with mechanism to detect congestion in the switch and return a binary signal B indicating congestion or no congestion. The flow control mechanism uses B and other network related information to generate a probability transmission table against which received packets are tested to determine proactively whether a packet is to be discarded or forwarded.

Abstract: An embedded processor complex contains multiple protocol processor units (PPUs). Each unit includes at least one, and preferably two independently functioning core language processors (CLPs). Each CLP supports dual threads thread which interact through logical coprocessor execution or data interfaces with a plurality of special purpose coprocessors that serve each PPU. Operating instructions enable the PPU to identify long and short latency events and to control and shift priority for thread execution based on this identification. The instructions also enable the conditional execution of specific coprocessor operations upon the occurrence or non occurrence of certain specified events.

Abstract: A network processor useful in network switch apparatus and methods of operating such a processor in which data flow handling and flexibility is enhanced by the cooperation of an embedded processor complex with a suite of peripherals, all formed on a common semiconductor substrate. The interface processors provide data paths for inbound and outbound data flow and operate under the control of instructions stored in an instruction store formed on the semiconductor substrate, while storage of transiting data flow portions is provided by memory peripherals and interfaces to external memory elements.

Abstract: An instruction memory system is shared by a plurality of processors and the system utilizes an increased bandwidth to support the combined number of processors. The total instruction address space is divided into code segments according to the disjoint tasks to be performed. The instruction codes of each processor are consolidated into one copy for control instructions and duplicate copies for other disjoint tasks such as inbound requests and outbound requests that have greater processor contention. Interleaving of the memory arrays for certain disjoint tasks serves to provide a larger number of instructions for these tasks. The system utilizes arbiters to receive all disjoint tasks and to control multiplexors that send addresses to memory arrays.

Abstract: A method and system are disclosed for allocating data input bandwidth from a source link to a plurality of N data queues each having a variable occupancy value, Qi(t), and a constant decrement rate, Di, where i designated the ith queue among the N queues. First, a threshold occupancy value, T, is designated for the N queues. During each time step of a repeating time interval, &Dgr;t, the occupancy value, Qi, is compared with T. In response to each and every of said N data queues having occupancy values exceeding T, pausing data transmission from the source link to the N data queues, such that overflow within the data queues is minimized. In response to at least one of the N data queues having an occupancy value less than or equal to T, selecting one among the N data queues to be incremented, and incrementing the selected data queue, such that underflow of the selected queue is minimized. In the context of scheduling one cell per time step, the value of T is one.

Abstract: A Resource Reservation System includes a Token Generation Unit (TGU) which generates and circulates among nodes of a communications system a Slotted Token (SLT) message having sub-fields to carry identification number for each input port in a node and the resource available for each input port. On receiving the message the Resource Control Unit (RCU) in each node can write port identification number, available resource in appropriate sub-fields of the SLT message, and reserve resources in other nodes by adjusting information in the sub-field associated with the other nodes.