Abstract: A peer-vector machine includes a host processor and a hardwired pipeline accelerator. The host processor executes a program, and, in response to the program, generates host data, and the pipeline accelerator generates pipeline data from the host data. Alternatively, the pipeline accelerator generates the pipeline data, and the host processor generates the host data from the pipeline data. Because the peer-vector machine includes both a processor and a pipeline accelerator, it can often process data more efficiently than a machine that includes only processors or only accelerators. For example, one can design the peer-vector machine so that the host processor performs decision-making and non-mathematically intensive operations and the accelerator performs non-decision-making and mathematically intensive operations.

Abstract: The present invention is generally directed to method and apparatus for emulating a portion of a stack. Certain embodiments of the invention manage a plurality of processor registers to store the top portion of the stack. Data is managed in these registers by managing a pointer that points to a current top-of-stack register. As data is pushed or popped from the stack, the top-of-stack point is incremented or decremented accordingly.

Abstract: A processor communication register (PCR) contained in each processor within a multiprocessor cluster network provides enhanced processor communication. Each PCR stores identical processor communication information that is useful in pipelined or parallel multi-processing. Each processor has exclusive rights to store to a sector within each PCR within the cluster network and has continuous access to read the contents of its own PCR. Each processor updates its exclusive sector within all of the PCRs via a private protocol or dedicated wireless network, instantly allowing all of the other processors within the cluster network to see the change within the PCR data, and bypassing the cache subsystem.

Abstract: A method and device to copy data fields from one or more source packets to one or more result packets. In a SET function, adjacent data fields in a source packet is copied to respective destination data fields in a result packet governed by a field locator packet. In an ESET function, data fields in respective source packets are copied to adjacent data fields in a result packet governed by a field locator packet. In an EXTRACT function, data fields in a source packet are copied to adjacent data fields in a result packet governed by a field locator packet. In a SCATTER function, adjacent data fields in a source packet are copied to data fields in respective result packets governed by a field locator packet.

Abstract: An accelerator 120 is tightly coupled to the normal execution unit 110. The operand store, which could be a register file 130, a stack based operand store or other operand store is shared by the execution unit and the accelerator unit. Operands may also be accessed as immediate values within the instructions themselves. The sequences of individual program instructions corresponding to computational subgraphs remain within a program but can be recognized by the accelerator as suitable for acceleration and when encountered are executed by the accelerator instead of by the normal execution unit. Within such tightly coupled arrangement problems can arise due to a lack of register resources within the system. The present technique provides that at least some intermediate operand values which are generated within the accelerator, but are determined not to be referenced outside of the computational subgraph concerned, are not written to the operand store.

Abstract: A data processing system 2 is provided which supports shift-and-insert instructions SLI, SRI which serve to shift a source data value by a specified shift amount and then insert bits from that shifted value other than the shifted-in bits into a destination value with the remaining bits within that destination value being unaltered.

Abstract: A method of executing program instructions may include receiving, in a processor, an instruction that causes the processor to read data from or write data to a portion of memory that is shared by one or more processes, at least one process of which manipulates data in a format that is different than a format of data in the shared portion of memory. The method may further include executing alternate instructions in place of the received instruction. The alternate instructions may effect transformation of data associated with the shared portion of memory from a first data format to a second data format.

Abstract: A register file bit includes a primary latch and a secondary latch with a feedback path and a context switch mechanism that allows a fast context switch when execution changes from one thread to the next. A bit value for a second thread of execution is stored in the primary latch, then transferred to the secondary latch. The bit value for a first thread of execution is then written to the primary latch. When a context switch is needed (when the first thread stalls and the second thread needs to begin execution), the register file bit can perform a context switch from the first thread to the second thread in a single clock cycle. The register file bit contains a backup latch inside the register file itself so that minimal extra wire paths are needed to or from the existing register file.

Abstract: A data processing system incorporates an instruction prefetch unit 8 including a static branch predictor 12. A static branch prediction cache 30, 32, 34 is provided for storing a most recently encountered static branch prediction such that a subsequent request to fetch the already encountered branch instruction can be identified before the opcode for that branch instruction is returned. The cached static branch prediction can thus redirect the prefetching to the branch target address sooner than the static predictor 12.

Abstract: The present invention is a method for implementing two architectures on a single chip. The method uses a fetch engine to retrieve instructions. If the instructions are macroinstructions, then it decodes the macroinstructions into microinstructions, and then bundles those microinstructions using a bundler, within an emulation engine. The bundles are issued in parallel and dispatched to the execution engine and contain pre-decode bits so that the execution engine treats them as microinstructions. Before being transferred to the execution engine, the instructions may be held in a buffer. The method also selects between bundled microinstructions from the emulation engine and native microinstructions coming directly from the fetch engine, by using a multiplexer or other means. Both native microinstructions and bundled microinstructions may be held in the buffer. The method also sends additional information to the execution engine.

Abstract: This invention pertains to apparatus, method and a computer program stored on a computer readable medium. The computer program includes instructions for use with an instruction unit having a code page, and has computer program code for partitioning the code page into at least two sections for storing in a first section thereof a plurality of instruction words and, in association with at least one instruction word, for storing in a second section thereof an extension to each instruction word in the first section. The computer program further includes computer program code for setting a state of at least one page table entry bit for indicating, on a code page by code page basis, whether the code page is partitioned into the first and second sections for storing instruction words and their extensions, or whether the code page is comprised instead of a single section storing only instruction words.

Abstract: A number of architectural and implementation approaches are described for using extra path (Epath) storage that operate in conjunction with a compute register file to obtain increased instruction level parallelism that more flexibly addresses the requirements of high performance algorithms. A processor that supports a single load data to a register file operation can be doubled in load capability through the use of an extra path storage, an additional independently addressable data memory path, and instruction decode information that specifies two independently load data operations. By allowing the extra path storage to be accessible by arithmetic facilities, the increased data bandwidth can be fully utilized.

Abstract: The data processing device and electronic equipment of the present invention perform pipeline control and include a fetch circuit which fetches instruction codes of a plurality of instructions in instruction queues. A prefix instruction decoder circuit performs a decode processing only on a prefix instruction. The prefix instruction decoder circuit receives the instruction code before decoding, judges whether or not the instruction is a given prefix instruction, and causes a target instruction to modify an information register to store information necessary for decoding a target instruction when the instruction is the given prefix instruction. A decoder circuit receives each of the instruction codes of the instructions other than the prefix instruction as a decode instruction and decodes the decode instruction. When the decode instruction is a target instruction, the target instruction modified by the prefix instruction is decoded based on the target instruction modifying information.

Abstract: A method of generating an expected TOS during translation of instructions. The method includes translating a first block of instructions executable in a first processor architecture, into a translated first block of instructions executable in a second processor architecture, wherein the translated first block of instructions operate with a stack of data entry positions. During the translation, an expected Top of Stack (TOS) position in the stack for the first block of code is generated.

Abstract: There is disclosed a data processor having a clustered architecture that comprises at least one branching cluster, at least one non-branching cluster and remote conditional branching control circuitry. Each of the clusters is capable of computing branch conditions, though only the branching cluster is operable to perform branch address computations. The remote conditional branching control circuitry, which is associated with each of the clusters, is operable in response to sensing a conditional branch instruction in a non-branching cluster to (i) cause the branching cluster to compute a branch address and a next program counter address,(ii) cause the non-branching cluster to compute a branch condition, and (iii) communicate the computed branch condition from the non-branching cluster to the branching cluster. The data processor then uses the computed branch condition to select one of the branch address or the next program counter address.

Abstract: A system and method of processing multiple swap requests including receiving a first swap request in a pipeline and executing the first swap request. A second swap request is also received in the pipeline immediately following the first swap request. The first swap request and the second swap request are examined to determine if the first swap request and the second swap request swap a same register.

Abstract: The present invention provides for a method and apparatus for the detection and prevention of and recovery from bogus branch predictions in a microprocessor. Micro-ops, decoded from a macro instruction, are stored in a decoded micro-op cache. Branch prediction logic determines whether a branch is bogus or not. If the branch taken was determined to be bogus, the present invention causes the micro-ops which descend from the original bogus branch micro-op instruction to be flagged and subsequently moved to the back-end of the processor for retirement. Further, the branch prediction logic (the branch prediction logic storage buffer) is updated as to what the actual direction of the branch was. In this manner then, bogus branches are detected, recovered from and further prevented.

Abstract: A device (1) to control processing of data elements (data_i), in which a thread is assigned to each data element (data_i), comprises a first unit (CS), which, during a first cycle, fetches an instruction (cs_ir_s) that is entered in the context of the thread assigned to the incoming data element (data_i), a second unit (IF), which, during a second cycle, fetches an instruction (if_ir_s) that succeeds a stipulated instruction in a stipulated thread, and a third unit (ID), which, during the second cycle, decodes the instruction prescribed for processing of the data element (data_i) and generates a data element processing signal (dec_o).

Abstract: For use in a wide-issue processor, a mechanism for, and method of, conditionally executing instructions and a digital signal processor (DSP) incorporating the mechanism or the method. In one embodiment, the mechanism includes: (1) a conditional execution block state machine that tags and generates link pointers for instructions located in a conditional execution block and (2) conditional link pointer registers, associated with stages in a pipeline of the processor, that contain and cause the link pointers to move therethrough as the instructions located in the conditional execution block move through the stages.

Abstract: Generally, the present invention provides a system and method for processing instructions of a computer program and for indicating instruction attribute and/or status information so that the efficiency of the processing system may be increased. In architecture, the system of the present invention utilizes a pipeline, a scoreboard, and hazard detection circuitry. The pipeline processes and executes instructions of a computer program. Many of the instructions include register identifiers that identify registers where data should be written when the instructions are executed. When the data produced by execution of one of the instructions has yet to be written to the register identified by the one instruction's register identifier and is unavailable for use in executing other instructions of the program, the one instruction's register identifier is transmitted to the scoreboard.