194 results about "Systolic array" patented technology

A method for routing and switching data packets from one or more incoming links to one or more outgoing links of a router. The method comprises receiving a data packet from the incoming link, assigning at least one outgoing link to the data packet based on the destination address of the data packet, and after the assigning operation, storing the data packet in a switching memory based on the assigned outgoing link. The data packet extracted from the switching memory, and transmitted along the assigned outgoing link. The router may include a network processing unit having one or more systolic array pipelines for performing the assigning operation.

A mechanism synchronizes among processors of a processing engine in an intermediate network station. The processing engine is configured as a systolic array having a plurality of processors arrayed as rows and columns. The mechanism comprises a barrier synchronization mechanism that enables synchronization among processors of a column (i.e., different rows) of the systolic array. That is, the barrier synchronization function allows all participating processors within a column to reach a common point within their instruction code sequences before any of the processors proceed.

A processor for use in a router, the processor having a systolic array pipeline for processing data packets to determine to which output port of the router the data packet should be routed. In one embodiment, the systolic array pipeline includes a plurality of programmable functional units and register files arranged sequentially as stages, for processing packet contexts (which contain the packet's destination address) to perform operations, under programmatic control, to determine the destination port of the router for the packet. A single stage of the systolic array may contain a register file and one or more functional units such as adders, shifters, logical units, etc., for performing, in one example, very long instruction word (vliw) operations. The processor may also include a forwarding table memory, on-chip, for storing routing information, and a cross bar selectively connecting the stages of the systolic array with the forwarding table memory.

This invention first presents SRAM based pipeline IP lookup architectures including an SRAM based systolic array architecture that utilizes multi-pipeline parallelism idea and elaborates on it as the base architecture highlighting its advantages. In this base architecture a multitude of intersecting and different length pipelines are constructed on a two dimensional array of processing elements in a circular fashion. The architecture supports the use of any type of prefix tree instead of conventional binary prefix tree. The invention secondly proposes a novel use of an alternative and more advantageous prefix tree based on binomial spanning tree to achieve a substantial performance increase. The new approach, enhanced with other extensions including four-side input and three-pointer implementations, considerably increases the parallelism and search capability of the base architecture and provides a much higher throughput than all existing IP lookup approaches making, for example, a 7 Tbps router IP lookup front end speed possible. Although theoretical worst-case lookup delay in this systolic array structure is high, the average delay is quite low, large delays being observed only rarely. The structure in its new form is scalable in terms of processing elements and is also well suited for the IPv6 addressing scheme.

A more computationally efficient and scalable systolic architecture is provided for computing the discrete Fourier transform. The systolic architecture also provides a method for reducing the array area by limiting the number of complex multipliers. In one embodiment, the design improvement is achieved by taking advantage of a more efficient computation scheme based on symmetries in the Fourier transform coefficient matrix and the radix-4 butterfly. The resulting design provides an array comprised of a plurality of smaller base-4 matrices that can simply be added or removed to provide scalability of the design for applications involving different transform lengths to be calculated. In this embodiment, the systolic array size provides greater flexibility because it can be applied for use with any transform length which is an integer multiple of sixteen.

A method and apparatus provide a systolic equalizer for Viterbi equalization of an 8-PSK signal distorted by passage through a communication channel. The systolic equalizer architecture is scalable to process, as examples, four, eight and 16 state received signals. An equalizer in accordance with this invention includes a logical arrangement of a plurality of instantiations of locally coupled processing elements forming a systolic array for processing in common received signal samples having distortion induced by passage through a communication channel. The equalizer outputs soft values for input to a decoder, the soft values representing an approximation of maximum a posteriori (MAP) probabilities. A trellis search procedure is employed to reconstruct estimates of a received signal sequence based on a reduced number of states. The reduced number of states is represented by a plurality of groups determined by partitioning a symbol constellation such that there are fewer groups than possible symbols.

Disclosed is a one-dimensional MFA systolic array for matrix computation using an MFA (modified Faddeeva algorithm), in which downward square MFA array processing and upward square MFA array processing are mapped to a one-dimensional array in horizontal directions, respectively. In each PE in the one-dimensional array, downward and upward MFA matrix calculations for two threads are executed. An input and an output are provided for each of PEs at both ends of the one-dimensional array.