Abstract: A high level programming language provides extensible data parallel semantics. User code specifies hardware and software resources for executing data parallel code using a compute device object and a resource view object. The user code uses the objects and semantic metadata to allow execution by new and/or updated types of compute nodes and new and/or updated types of runtime libraries. The extensible data parallel semantics allow the user code to be executed by the new and/or updated types of compute nodes and runtime libraries.

Abstract: A high level programming language provides an agile communication operator that generates a segmented computational space based on a resource map for distributing the computational space across compute nodes. The agile communication operator decomposes the computational space into segments, causes the segments to be assigned to compute nodes, and allows the user to centrally manage and automate movement of the segments between the compute nodes. The segment movement may be managed using either a full global-view representation or a local-global-view representation of the segments.

Abstract: A multi-core processor includes microcode distributed in each core enabling each core to participate in a de-centralized inter-core state discovery process. In a related microcode-implemented method, states of a multi-core processor are discovered by at least two cores participating in a de-centralized inter-core state discovery process. The inter-core state discovery process is carried out through a combination of microcode executing on each participating core and signals exchanged between the cores through sideband non-system-bus communication wires. The discovery process is unmediated by any centralized non-core logic. Applicable discoverable states include target and composite power states, whether and how many cores are enabled, the availability and distribution of various resources, and hierarchical structures and coordination systems for the cores.

Abstract: Systems and methods are provided for managing access to registers. A system may include a set of direct registers and a set of indirect registers. The indirect registers may be accessed through the direct registers, and the direct registers may provide various features to provide faster access to the indirect registers. One of the direct registers may indicate access modes for accessing the indirect registers. The access modes may include auto-increment, auto-decrement, auto-reset, and no change modes. Based on the access mode, the currently accessed address may be automatically modified after accessing the indirect register at the address.

Abstract: Techniques are described to provide the capability to halt execution of a thread by a processor and potentially lower power consumption of the processor while responding to events in a timely manner. An operating system provided system call allows for identification of events that cause the execution of the thread to resume. A processor core uses a signal mask and translation unit that monitors for the identified events. In the event an event is detected, the thread unhalts and determines a manner to process the event.

Abstract: An apparatus for executing fixed width instructions in a multiple execution unit system has a device for fetching instructions from a memory, and a decoder for decoding each fetched instruction in turn. A determination is made as to whether each decoded instruction includes a portion to fetch a locally stored instruction from a local store. If it does, the locally stored instruction is fetched and locally stored portion are executed.

Abstract: Systems and methods that allow for extracting a field from data stored in a pair of registers using two instructions. A first instruction extracts any part of the field from a first register designated as a first source register, and executes a second instruction extracting any part of the field from a second general register designated as a second source register. The second instruction inserts any extracted field parts in a result register.

Abstract: Different instruction sets are provided for different units of execution such as threads, processes, and execution contexts. Execution units may be associated with instruction sets. The instruction sets may have mutually exclusive opcodes, meaning an opcode in one instruction set is not included in any other instruction set. When executing a given execution unit, the processor only allows execution of instructions in the instruction set that corresponds to the current execution unit. A failure occurs if the execution unit attempts to directly execute an instruction in another instruction set.

Abstract: A programming language may include hint instructions that may notify a programming idiom accelerator that a programming idiom is coming. An idiom begin hint exposes the programming idiom to the programming idiom accelerator. Thus, the programming idiom accelerator need not perform pattern matching or other forms of analysis to recognize a sequence of instructions. Rather, the programmer may insert idiom hint instructions, such as an idiom begin hint, to expose the idiom to the programming idiom accelerator. Similarly, an idiom end hint may mark the end of the programming idiom.

Abstract: A computer-implemented method and a system for computational acceleration of seismic data processing are described. The method includes defining a specific non-uniform memory access (NUMA) scheduling for a plurality of cores in a processor according to data to be processed; and running two or more threads through each of the plurality of cores.

Abstract: Method, apparatus and system embodiments to schedule user-level OS-independent “shreds” without intervention of an operating system. For at least one embodiment, the shred is scheduled for execution by a scheduler routine rather than the operating system. The scheduler routine may receive compiler-generated hints from a compiler. The compiler hints may be generated by the compiler without user-provided pragmas, and may be passed to the scheduler routine via an API-like interface. The interface may include a scheduling hint data structure that is maintained by the compiler. Other embodiments are also described and claimed.

Abstract: A data processing apparatus includes an output unit. The output unit determines, when parallel control is performed in a data processor created in the data processing apparatus so that plural processing modules forming the data processor perform data processing in parallel, on the basis of a value representing a parallel-processing time for which at least two processing modules are operated in parallel and a value representing a control time, which is not necessary when serial control is performed so that the processing modules serially perform data processing but which is necessary when the parallel control is performed so that the processing modules perform data processing in parallel, whether a time necessary to complete data processing performed by the data processor under the parallel control would be shorter than a time necessary to complete data processing performed by the data processor under the serial control, and outputs a determination result.

Abstract: Programs can be optimized at runtime prior to execution to enhance performance. Program instructions/operations designated for execution can be recorded and subsequently optimized at runtime prior to execution, for instance by performing transformations on the instructions. For example, such optimization can remove, reorder, and/or combine instructions, among other things.

Abstract: A method of processing data comprising performing a sequence of operation instructions with variable operand size, wherein respective size codes for different source and destination operands are obtained and registered separately from performing the sequence of operation instructions, and the sequence of operation instructions is performed using operand sizes defined by the registered size codes, the operation instructions of the sequence not themselves containing size codes.

Abstract: Embodiments are directed to implementing a generic SIMD data type in software code. In an embodiment, a computer system accesses a portion of software code that includes an algorithm with a generic SIMD data type that includes a variable number of elements. The algorithm with the generic SIMD data type is to be processed by a specific processor that includes various specific hardware features. The computer system determines at runtime a portion of customized processor-specific code that is to be used with the specified processor based on the generic SIMD data type, wherein the runtime determination resolves the number of elements that are to be used with the specified processor. The computer system also processes the software code including the algorithm with the generic SIMD data type using the determined, customized processor-specific code.

Abstract: An apparatus and method are provided for performing re-arrangement operations on data. The data processing apparatus has a register data store with a plurality of registers for storing data, and processing logic for performing a sequence of operations on data including at least one re-arrangement operation. The processing logic has scalar processing logic for performing scalar operations and SIMD processing logic for performing SIMD operations. The SIMD processing logic is responsive to a re-arrangement instruction specifying a family of re-arrangement operations to perform a selected re-arrangement operation from that family on a plurality of data elements constituted by data in one or more registers identified by the re-arrangement instruction. The selected re-arrangement operation is dependent on at least one parameter provided by the scalar processing logic, that parameter identifying a data element width for the data elements on which the selected re-arrangement operation is performed.

Abstract: A system and method for successfully performing reduction operations in a multi-threaded SIMD (single-instruction multiple-data) system while one or more threads are disabled allows for the reduction operations to be performed without a performance penalty compared with performing the same operation with all of the threads enabled. The source data for each intermediate computation of the reduction operation is remapped by a configurable crossbar as needed to avoid using invalid data from the disabled threads. The remapping function is transparent to the user and enables correct execution of order invariant reduction operations and order dependent prefix-reduction operations.

Abstract: A system and method for efficiently performing bit-field extraction and bit-field combination operations in a processor is provided. The system includes a plurality of general purpose registers, a plurality of predicate registers, and at least one execution unit configured to extract a plurality of bit fields from a source reservoir and to populate a plurality of destination lanes in response to a single instruction. In addition, the execution unit is configured to write supplied fill data into the source reservoir if the number of bits in the source reservoir is less than a predetermined number. In addition or alternatively, the system may include at least one execution unit configured to combine a plurality of bit fields from a plurality of source lanes into a continuous bit stream in response to a single instruction executable by the processor.

Abstract: Disclosed are computer systems, a plurality of methods and a computer program for preventing a delay in execution time of one or more instructions. The computer system includes: a lock unit for executing an instruction to acquire exclusive-use of the external resource and an instruction to release the exclusive-use of the external resource in the one or more threads; a counter unit for increasing or decreasing a value of a corresponding one of counters respectively associated with the threads; and a controller for controlling an execution order of the instructions to be executed by exclusively using the external resource and instructions that causes a delay in the execution time of the instructions to be executed by exclusively using the external resource.

Abstract: One embodiment of the present invention sets forth a technique for efficiently performing voting operations within a multi-threaded parallel-processing system. A group of related parallel program threads executes within a processor core together in parallel. A new instruction, called a “vote” instruction, is introduced that enables a parallel program thread to post an individual vote within the context of the group of related threads and to receive the result of the vote. In this fashion, the vote instruction advantageously reduces overhead associated with inter-thread communication, thereby improving overall system performance.

Abstract: A method and apparatus is presented for identifying instructions in a stream of information by preprocessing the stream of information, creating a vector of instructions and breaking the vector of instructions into two or more vectors for picking the identified instructions at a high frequency.

Abstract: A method and apparatus for load step, or instantaneous current spike, mitigation are provided. In the method and apparatus, load steps are mitigated if a computer system a whole is lightly load, which may be determined by the power consumption of the computer system. Further, load steps are mitigated if a number of processor cores capable of inducing a load step is higher than a threshold. The Advanced Configuration and Power Interface (ACPI) performance state of the cores is used to determine a core's potential for generating a load step. A processor core is instructed to mitigate load steps if conditions are met for the mitigation.

Abstract: An information processing apparatus includes the following elements. A generator generates, on the basis of instruction information which describes processing to be executed for obtaining output data from raw data, processing definition information that defines details of the processing, upon inputting the instruction information. A determination unit determines whether output data associated with the currently generated processing definition information and data to be used as raw data is stored in a first memory, the first memory storing therein output data which has been obtained in accordance with previously generated processing definition information in association with data used as the raw data and the processing definition information. An output unit outputs, if the determination unit determines that the output data is stored in the first memory, the output data stored in the first memory without causing a processor to execute the processing.

Abstract: A pipeline processor architecture, processor, and methods are provided. In one implementation, a processor is provided that includes an instruction fetch unit operable to fetch instructions associated with a plurality of processor threads, a decoder responsive to the instruction fetch unit, issue logic responsive to the decoder, and a register file including a plurality of banks corresponding to the plurality of processor threads. Each bank is operable to store data associated with a corresponding processor thread. The processor can include a set of registers corresponding to each of a plurality of processor threads. Each register within a set is located either before or after a pipeline stage of the processor.

Abstract: A hardware device for concurrently processing a fixed set of predetermined tasks associated with an algorithm which includes a number of processes, some of the processes being dependent on binary decisions, includes a plurality of task units for processing data, making decisions and/or processing data and making decisions, including source task units and destination task units. A task interconnection logic means interconnect the task units for communicating actions from a source task unit to a destination task unit. Each of the task units includes a processor for executing only a particular single task of the fixed set of predetermined tasks associated with the algorithm in response to a received request action, and a status manager for handling the actions from the source task units and building the actions to be sent to the destination task units.

Abstract: A method for controlling the execution of at least one program in an electronic circuit and a processor for executing a program, in which at least one volatile memory area of the circuit is, prior to the execution of the program to be controlled, filled with first instructions resulting in an exception processing; the program contains instructions for replacing all or part of the first instructions with second valid instructions; and the area is called for execution of all or part of the instruction that it contains at the end of the execution of the instruction program.

Abstract: A hardware device for processing the tasks of an algorithm of the type having a number of processes the execution of some of which depend on binary decisions has a plurality of task units (10, 12, 14), each of which are associated with a task defined as being either one process or one decision or one process together with a following decision. A task interconnection logic block (16) is connected to each task unit for communicating actions from a source task unit to a destination task unit. Each task unit includes a processor (18) for processing the steps of the associated task when a received action requests such a processing. A status manager (20) handles actions coming from other task units and builds actions to be sent to other task units.

Abstract: A method and system for implementing a micro-sequence based security model in a processor. More particularly, micro-sequences and JSM hardware resources are employed to construct a security model invisible to applications, and when memory constraints are in place, extend a complex security model in JSM code by implementing a micro-sequence security trigger. The method includes micro-sequence based security policy that determines whether an instruction accesses a privileged resource associated with a processor and when not already in privilege mode and not executing a micro-sequence, the micro-sequence based security policy is applied to the instruction to control access to the privileged resource according to the security policy.

Abstract: Mechanisms for controlling rollover or reset of hardware performance counters in the data processing system. A signal indicating that a rollover or reset of a first hardware performance counter has occurred is received and it is determined if the first hardware performance counter is analytically related to one or more second hardware performance counters based on defined ganged hardware performance counter sets. A signal is sent to each of the one or more second hardware performance counters in response to a determination that the first hardware performance counter is analytically related to the one or more second hardware performance counters. Each of the one or more second hardware performance counters is reset to an initial value in response to the one or more second hardware performance counters receiving the signal from the ganged hardware performance counter rollover/reset logic.

Abstract: A data processing apparatus may include a processing unit that performs processing related to data, a first register that holds a value for defining an operation of the processing unit, a second register that holds a value output from the first register, the second register outputting the value to the processing unit, a first control unit that performs control for writing a value in the first register, a second control unit that performs control for rewriting the value held by the second register with the value output from the first register, after the value is written in the first register, and a third control unit that performs control for rewriting the value held by the second register with an invalid value, at which the processing of the processing unit is stopped, during a period for which the value is written in the first register.

Abstract: A hardwired core is embedded in an integrated circuit having programmable circuitry. The hardwired core has a microprocessor; a crossbar interconnect coupled to processor local buses of the microprocessor; and a memory controller interface coupled to the crossbar interconnect. The crossbar interconnect provides pipelines for coupling the hardwired core to the programmable circuitry. The microprocessor, the crossbar interconnect, and the memory controller interface are all capable of operating at a first frequency of operation, and the memory controller interface is further capable of being set to operate at a second frequency of operation having an integer ratio with respect to the first frequency of operation. The crossbar interconnect is configured to direct transactions initiated by the microprocessor to the memory controller interface for accessing one or more memory devices coupled to the memory controller interface via a memory controller.

Abstract: A system including a dual function adder is described. In one embodiment, the system includes an adder. The adder is configured for a first instruction to determine an address for a hardware prefetch if the first instruction is a hardware prefetch instruction. The adder is further configured for the first instruction to determine a value from an arithmetic operation if the first instruction is an arithmetic operation instruction.

Abstract: A method and apparatus are provided for executing instructions of a multi-threaded processor having multiple hardware threads (32, 34) with differing hardware resources comprising the steps of receiving a plurality of streams of instructions (38, 44) and determining which hardware threads are able to receive instructions for execution (40, 46), determining whether a thread determined to be available for executing an instructions has the hardware resources available required by that instructions (36) and executing the instruction in dependence on the result of the determination (50).

Abstract: A method for performing multiple-operand logical operations in a single instruction includes the steps of: generating a table defining a correspondence between a plurality of input variables to a multiple-operand logical operation and a plurality of output results of the multiple-operand logical operation; encoding the table to generate a set of values for use by the single instruction, each value being indicative of an output result of the multiple-operand logical operation as a function of a corresponding unique combination of values of the input variables; and at least one processor performing the multiple-operand logical operation in a single instruction as a function of the set of values for a prescribed combination of values of the input variables.

Abstract: One embodiment of the present invention provides a system that merges checkpoints on a processor. The system starts by executing instructions speculatively during a speculative-execution episode. The system then generates a first checkpoint and a second checkpoint during the speculative-execution episode. Next, the system merges the first checkpoint with the second checkpoint during the speculative-execution episode, wherein merging the first and second checkpoints conserves processor resources.

Abstract: Improved instruction set and core design, control and communication for programmable microprocessors is disclosed, involving the strategy for replacing centralized program sequencing in present-day and prior art processors with a novel distributed program sequencing wherein each functional unit has its own instruction fetch and decode block, and each functional unit has its own local memory for program storage; and wherein computational hardware execution units and memory units are flexibly pipelined as programmable embedded processors with reconfigurable pipeline stages of different order in response to varying application instruction sequences that establish different configurations and switching interconnections of the hardware units.

Abstract: The scheduling of multiple request to be processed by a number of deterministic finite automata-based graph thread engine (DTE) workstations is processed by a novel scheduler. The scheduler may select an entry from an instruction in a content search apparatus. Using attribute information from the selected entry, the scheduler may thereafter analyze a dynamic scheduling table to obtain placement information. The scheduler may determine an assignment of the entry, using the placement information, that may limit cache thrashing and head of line blocking occurrences. Each DTE workstation may including normalization capabilities. Additionally, the content searching apparatus may employ an address memory scheme that may prevent memory bottle neck issues.

Abstract: A method and apparatus for controlling system access to a memory that includes receiving first and second instructions, and evaluating whether both instructions can architecturally complete. When at least one instruction cannot architecturally complete, delaying both instructions. When both instructions can architecturally complete and at least one is a write instruction, adjusting a write control of the memory to account for an evaluation delay. The evaluation delay can be sufficient to evaluate whether both instructions can architecturally complete. The evaluation delay can be input to the write control and not the read control of the memory. A precharge clock of the memory can be adjusted to account for the evaluation delay. Evaluating whether both instructions can architecturally complete can include determining whether data for each instruction is located in a cache, and whether the instructions are memory access instructions.

Abstract: A compressed instruction processing device has: a compressed instruction expanding circuit which expands a compressed instruction code that include a difference code between an instruction code being a compression object and a reference instruction code and which outputs an expanded instruction code; an instruction buffer storing the instruction code expanded by the compressed instruction expanding circuit; and an execution section executing the instruction code expanded by the compressed instruction expanding circuit, wherein the compressed instruction expanding circuit outputs the expanded instruction code by inputting the instruction code in the instruction buffer as the reference instruction code and adding the reference instruction code and the difference code in the compressed instruction code.

Abstract: A Baseboard Management Controller (BMC) controlling method includes the steps of dividing a memory of a BMC into an original region and customized region, in which the original region includes at least one original sensor data record (SDR) and original platform event filter (PEF) corresponding to each other; providing an instruction set to at least one external system, in which the external system manages at least one customized SDR and customized PEF corresponding to each other in the customized region through the instruction set; polling the original SDR in the original region and the customized SDR in the customized region; determining whether values of the SDRs obtained through polling conform to a plurality of critical values individually corresponding to the SDRs; and obtaining a processing policy according to the corresponding PEF when at least one value of the SDR does not conform to the corresponding critical value.

Abstract: A method executed by an instruction set on a processor is described. The method includes providing a tbbit instruction, inputting a first index for the tbbit instruction, loading a second value for the tbbit instruction, wherein the second value comprises at least 2b bits, using selected b bits of the first index to select at least one target bit in the loaded second value, shifting the target bit into the bottom of the first index, and computing a second index based on the shifting of the target bit into the bottom of the first index. Other methods and variations are also described.

Abstract: A system, method and computer program product for managing a plurality of applications in a computer cluster. Each application is able to run on a particular node in the cluster. In one embodiment, associations are maintained among a plurality of modes and the plurality of applications, with each application being associated with at least one mode. Responsive to designation of at least one mode as active for the cluster, each application that is associated with an active mode is flagged as eligible for activation, each inactive application that is not associated with any active mode is flagged as ineligible for activation, and each active application that is not associated with any active mode is flagged as ineligible for activation and inactivated. Flagging as eligible, flagging as ineligible and flagging as ineligible and inactivating may be performed in any order, and inactivating is sequenced according to dependencies among the applications.

Abstract: Disclosed are selected embodiments of a processor that may include a plurality of thread units and a register file architecture to support speculative multithreading. For at least one embodiment, live-in values for a speculative thread are computed via execution of a precomputation slice and are stored in a validation buffer for later validation. A global register file holds the committed architecture state generated by a non-speculative thread. Each thread unit includes a local register file. A directory indicates, for each architectural register, which speculative thread(s) have generated a value for the architectural register. Other embodiments are also described and claimed.

Abstract: A generator of a signal including a memory in which instructions are stored, each instruction including a code portion and an argument portion; circuitry for successively reading instructions stored in the memory; decoding circuitry capable of receiving, for each read instruction, the code portion of the instruction and of providing an activation signal which depends on the code portion; and circuitry for providing the signal capable of receiving, for each read instruction, the argument portion of the instruction and capable, according to the activation signal, of storing the argument portion and of providing the signal equal to the argument portion or of providing the signal equal to the previously-stored argument portion.

Abstract: An emulation processing method causing a computer including a first and a second processor to execute emulation processing, the emulation processing method includes: calculate a next instruction address next to a received instruction address, and transmit, to the second processor, the calculated instruction address and instruction information read out on the basis of the calculated instruction address, transmit, to the first processor, a first instruction address that is an instruction address included in an execution result of executed processing, and execute processing based on the instruction information received from the first processor, when a second instruction address that is the instruction address received from the first processor is identical to the first instruction address, and read out instruction information on the basis of the first instruction address and execute processing based on the instruction information read out, when the second instruction address is not identical to the first instruct

Abstract: A high level programming language provides a tile communication operator that decomposes a computational space into sub-spaces (i.e., tiles) that may be mapped to execution structures (e.g., thread groups) of data parallel compute nodes. An indexable type with a rank and element type defines the computational space. For an input indexable type, the tile communication operator produces an output indexable type with the same rank as the input indexable type and an element type that is a tile of the input indexable type. The output indexable type provides a local view structure of the computational space that enables coalescing of global memory accesses in a data parallel compute node.

Abstract: The present invention includes a method for obtaining dynamic operating parameter information of a semiconductor device such as a processor, determining dynamic usage of the device, either as a whole or for one or more portions thereof, based on the dynamic operating parameter information, and dynamically estimating a remaining lifetime of the device based on the dynamic usage. Depending on the estimated remaining lifetime, the device may be controlled in a desired manner.

Abstract: A system is presented that is configured to reduce power consumption when performing processing tasks. The system includes a first processing entity capable of performing a set of operations, and a second processing entity configured to consume less power than the first processing entity and capable of performing a subset of operations that is part of the set of operations. During system operation, the second processing entity is configured to perform the subset of operations instead of the first processing entity.

Abstract: The present invention provides software programmable hardware state machines to detect a cause of an error in a processor and prevent the error from occurring. In example embodiments, processors, systems and methods are provided to prevent an unwanted change in architectural state from occurring as a result of execution of a specific sequence of instruction types. A processor core is provided that includes an execution unit, a programmable mask register and a buffer that stores values representing instructions dispatched to the execution unit. The processor core also includes control logic to determine whether there is a match between a sequence in the mask register and a sequence in the buffer and, upon detecting a match, to generate control signals to perform a desired action. The desired action prevents an unwanted change from occurring to the architectural state of the processor.

Abstract: An apparatus and method is described herein for conditionally committing /andor speculative checkpointing transactions, which potentially results in dynamic resizing of transactions. During dynamic optimization of binary code, transactions are inserted to provide memory ordering safeguards, which enables a dynamic optimizer to more aggressively optimize code. And the conditional commit enables efficient execution of the dynamic optimization code, while attempting to prevent transactions from running out of hardware resources. While the speculative checkpoints enable quick and efficient recovery upon abort of a transaction. Processor hardware is adapted to support dynamic resizing of the transactions, such as including decoders that recognize a conditional commit instruction, a speculative checkpoint instruction, or both. And processor hardware is further adapted to perform operations to support conditional commit or speculative checkpointing in response to decoding such instructions.