Abstract: The present invention describes a method and system for an interface for integrating reconfigurable processors into a general purpose computing system. In particular, the system resides in a computer system containing standard instruction processors, as well as reconfigurable processors. The interface includes a command processor, a command list memory, various registers, a direct memory access engine, a translation look-aside buffer, a dedicated section of common memory, and a dedicated memory. The interface is controlled via commands from a command list that is created during compilation of a user application, or various direct commands.

Abstract: The present invention describes a method and system for an interface for integrating reconfigurable processors into a general purpose computing system. In particular, the system resides in a computer system containing standard instruction processors, as well as reconfigurable processors. The interface includes a command processor, a command list memory, various registers, a direct memory access engine, a translation look-aside buffer, a dedicated section of common memory, and a dedicated memory. The interface is controlled via commands from a command list that is created during compilation of a user application, or various direct commands.

Abstract: An embodiment of the invention includes a method of simulating a hybrid instruction processor and reconfigurable processor implemented algorithm which utilizes a runtime selectable emulation library that emulates a reconfigurable processor and its resources, and a control-data flow emulator that emulates the reconfigurable logic for the algorithm. Another embodiment of the invention includes a method of simulating a control-dataflow graph that includes building an internal representation of the control-dataflow graph that includes one or more dataflow code blocks, and simulating the control-dataflow graph as a sequence of code block dataflow executions, where control is passed from one code block to another code block based on the output value of the code block until EXIT is reached.

Abstract: A reconfigurable processor that includes a computational unit and a data prefetch unit coupled to the computational unit, where the data prefetch unit retrieves data from a memory and supplies the data to the computational unit through memory and a data access unit, and where the data prefetch unit, memory, and data access unit is configured by a program. Also, a reconfigurable hardware system that includes a common memory; and one or more reconfigurable processors coupled to the common memory, where at least one of the reconfigurable processors includes a data prefetch unit to read and write data between the unit and the common memory, and where the data prefetch unit is configured by a program executed on the system. In addition, a method of transferring data that includes transferring data between a memory and a data prefetch unit in a reconfigurable processor; and transferring the data between a computational unit and the data prefetch unit.

Abstract: A control-flow dataflow graph pipelined loop structure that includes a loop body that processes an input value to generate an output value in successive iterations of the loop body, where the output value is captured by a circulate node coupled to the loop body, a loop valid node coupled to the loop body that determines a final loop iteration, and an output value storage node coupled to the circulate node, where the output value storage node ignores output values generated after the loop valid node determines the final loop iteration has occurred. Also, a control-flow dataflow graph pipelined loop structure that includes a loop body that processes an input value to generate an output value in successive iterations of the loop body, where the output value is captured by a circulate node coupled to the loop body, and a loop driver node coupled to the circulate node, where the loop driver node sets a period for each iteration of the loop body.

Abstract: Embodiments of the invention include a mechanism for explicit communication in a clustered multiprocessor system that supports low-latency, protected, user-mode, communication across the machine boundaries of a clustered multiprocessor. Data transport may be accomplished over persistent, unidirectional, point-to point connections, each of which may be embodied in a small amount of state at each end, along with a statically allocated per-connection memory buffer, which may be directly accessible to the transport mechanism at both ends of each notional link. System Memory protection may be afforded by operating system (“OS”) facilitated allocation of both restricted control of the network interface, and responsibility for data transport, to an application thread that may execute in the context of the processor-managed virtual address space. Connection buffer protection may be afforded by restricting access to connection state to those entries associated with the currently controlling thread.

Abstract: A computer system architecture and memory controller for close-coupling within a hybrid computing system using an adaptive processor interface port (“APIP”) added to, or in conjunction with, the memory and I/O controller chip of the core logic. Memory accesses to and from this port, as well as the main microprocessor bus, are then arbitrated by the memory control circuitry forming a portion of the controller chip. In this fashion, both the microprocessors and the adaptive processors of the hybrid computing system exhibit equal memory bandwidth and latency. In addition, because it is a separate electrical port from the microprocessor bus, the APIP is not required to comply with, and participate in, all FSB protocol. This results in reduced protocol overhead which results higher yielded payload on the interface.

Abstract: A computing system having at least one microprocessor and a memory subsystem coupled to the at least one microprocessor. A memory controller is coupled to manage memory transactions between the memory subsystem and the at least one microprocessor. At least one arbitration port is coupled to the memory controller and configured to receive an external arbitration signal.

Abstract: A system and method for compiling computer code written to conform to a high-level language standard to generate a unified executable containing the hardware logic for a reconfigurable processor, the instructions for a traditional processor (instruction processor), and the associated support code for managing execution on a hybrid hardware platform. Explicit knowledge of writing hardware-level design code is not required since the problem can be represented in a high-level language syntax. A top-level driver invokes a standard-conforming compiler that provides syntactic and semantic analysis. The driver invokes a compilation phase that translates the CFG representation being generated into a hybrid controlflow-dataflow graph representation representing optimized pipelined logic which may be processed into a hardware description representation.

Abstract: A multi-adaptive processor element architecture incorporating a field programmable gate array (“FPGA”) control element having at least one embedded processor core and a pair of user FPGAs forming a user array is disclosed in conjunction with high volume dynamic random access memory (“DRAM”) and dual-ported static random access memory (“SRAM”) banks. In operation, the DRAM is “read” using its fast sequential burst modes and the lower capacity SRAM banks are then randomly loaded allowing the user FPGAs to experience very high random access data rates from what appears to be a very large virtual SRAM. The reverse also happens when the user FPGAs are “writing” data to the SRAM banks. These overall control functions may be managed by an on-chip DMA engine that is implemented in the control FPGA.

Abstract: An embodiment of the invention includes a system for partitioning a control-flow graph representation into a reconfigurable portion and an instruction processor portion. Another embodiment of the invention includes a method of partitioning a control-dataflow graph representation that includes dividing the control-dataflow graph into two or more partition blocks, comparing the estimated performance of at least one of the partition blocks as reconfigurable logic versus instruction processor code; and assigning said at least one of the partition blocks to reconfigurable hardware or an instruction processor based on said comparing step.

Abstract: A multiprocessor computer architecture incorporating a plurality of programmable hardware memory algorithm processors (“MAP”) in the memory subsystem. The MAP may comprise one or more field programmable gate arrays (“FPGAs”) which function to perform identified algorithms in conjunction with, and tightly coupled to, a microprocessor and each MAP is globally accessible by all of the system processors for the purpose of executing user definable algorithms. A circuit within the MAP signals when the last operand has completed its flow thereby allowing a given process to be interrupted and thereafter restarted. Through the use of read only memory (“ROM”) located adjacent the FPGA, a user program may use a single command to select one of several possible pre-loaded algorithms thereby decreasing system reconfiguration time.

Abstract: The present invention includes a method of computing a function array in reconfigurable hardware that includes forming in the reconfigurable hardware a first delay queue and a second delay queue, inputting from a source array outside the reconfigurable hardware a first value into the first delay queue and a second value into the second delay queue, defining in the reconfigurable hardware a window array comprising a first cell and a second cell, inputting the first value from the first delay queue into the first cell and the second value from the second delay queue into the second cell, and calculating an output value for the function array based on the window array. The present invention also includes a method of loop stripmining and a method of calculating output values in a fused producer/consumer loop structure.

Abstract: A system and method for enhancing the utilization of available bandwidth for an uncached device. Data written to the device is done so by striding the available data into multiple data elements of the appropriate size for the uncached device. Data read from the device is retrieved from multiple addresses on the uncached device to avoid unnecessary waits cycles in the processor.

Abstract: An enhanced switch/network adapter port (“SNAP™”) including collocated shared memory resources (“SNAPM™”) in a dual in-line memory module (“DIMM”) or any other memory module format for clustered computing systems employing direct execution logic such as multi-adaptive processor elements (“MAP®”, all trademarks of SRC Computers, Inc.). Functionally, the SNAPM modules incorporate and properly allocate memory resources so that the memory appears to the associated dense logic device(s) (e.g. a microprocessor) to be functionally like any other system memory such that no time penalties are incurred when accessing it.

Abstract: A method and apparatus including a plurality of data processing units. A plurality of memory banks having a shared address space are coupled to the processors by a crossbar coupling to enable reading and writing data between the processors and memory banks. A unidirectional network couples the memory banks and the processors to enable cache coherency messages to be transmitted from the memory to the processors. A plurality of semaphore registers are implemented within the shared address space of the memory banks wherein the semaphore registers are accessible by the processors through the crossbar coupling.

Abstract: A system and method for accelerating web site access and processing utilizing a multiprocessor computer system incorporating reconfigurable and standard microprocessors as the web site server. One or more reconfigurable processors may be utilized, for example, in accelerating site visitor demographic data processing, real time web site content updating, database searches and other processing associated with e-commerce applications. In a particular embodiment disclosed, all of the reconfigurable and standard microprocessors may be controlled by a single system image of the operating system, although cluster management software may be utilized to cause a cluster of microprocessors to appear to the user as a single copy of the operating system.

Abstract: A cache coherency directory for a shared memory multiprocessor computer system. A data structure is associated with each cacheable memory location, the data structure comprising locations for storing state values indicating an exclusive state, a shared state, an uncached state, a busy state, a busy uncached state, a locked state, and a pending state. The busy state and pending state cooperate to reserve a cache line for future use by a processor while the cache line is currently being used by one or more other processors.

Abstract: An enhanced memory algorithmic processor (“MAP”) architecture for multiprocessor computer systems comprises an assembly that may comprise, for example, field programmable gate arrays (“FPGAs”) functioning as the memory algorithmic processors. The MAP elements may further include an operand storage, intelligent address generation, on board function libraries, result storage and multiple input/output (“I/O”) ports. The MAP elements are intended to augment, not necessarily replace, the high performance microprocessors in the system and, in a particular embodiment of the present invention, they may be connected through the memory subsystem of the computer system resulting in it being very tightly coupled to the system as well as being globally accessible from any processor in a multiprocessor computer system.

Abstract: A split directory-based cache coherency technique utilizes a secondary directory in memory to implement a bit mask used to indicate when more than one processor cache in a multi-processor computer system contains the same line of memory which thereby reduces the searches required to perform the coherency operations and the overall size of the memory needed to support the coherency system. The technique includes the attachment of a “coherency tag” to a line of memory so that its status can be tracked without having to read each processor's cache to see if the line of memory is contained within that cache. In this manner, only relatively short cache coherency commands need be transmitted across the communication network (which may comprise a Sebring ring) instead of across the main data path bus thus freeing the main bus from being slowed down by cache coherency data transmissions while removing the bandwidth limitations inherent in other cache coherency techniques.