Abstract: A processor comprises a plurality of arithmetic units, an accumulator unit, and a reduction unit coupled between the plurality of arithmetic units and the accumulator unit. The reduction unit receives products of vector elements from the arithmetic units and a first accumulator value from the accumulator unit, and processes the products and the first accumulator value to generate a second accumulator value for delivery to the accumulator unit. The processor implements a plurality of vector multiply and reduce operations having guaranteed sequential semantics, that is, operations which guarantee that the computational result will be the same as that which would be produced using a corresponding sequence of individual instructions.

Abstract: A multithreaded processor comprises a plurality of hardware thread units, an instruction decoder coupled to the thread units for decoding instructions received therefrom, and a plurality of execution units for executing the decoded instructions. The multithreaded processor is configured for controlling an instruction issuance sequence for threads associated with respective ones of the hardware thread units. On a given processor clock cycle, only a designated one of the threads is permitted to issue one or more instructions, but the designated thread that is permitted to issue instructions varies over a plurality of clock cycles in accordance with the instruction issuance sequence. The instructions are pipelined in a manner which permits at least a given one of the threads to support multiple concurrent instruction pipelines.

Abstract: A multithreaded processor comprises a plurality of hardware thread units, an instruction decoder coupled to the thread units for decoding instructions received therefrom, and a plurality of execution units for executing the decoded instructions. The multithreaded processor is configured for controlling an instruction issuance sequence for threads associated with respective ones of the hardware thread units. On a given processor clock cycle, only a designated one of the threads is permitted to issue one or more instructions, but the designated thread that is permitted to issue instructions varies over a plurality of clock cycles in accordance with the instruction issuance sequence. The instructions are pipelined in a manner which permits at least a given one of the threads to support multiple concurrent instruction pipelines.

Abstract: A multithreaded processor comprises a plurality of hardware thread units, an instruction decoder coupled to the thread units for decoding instructions received therefrom, and a plurality of execution units for executing the decoded instructions. The multithreaded processor is configured for controlling an instruction issuance sequence for threads associated with respective ones of the hardware thread units. On a given processor clock cycle, only a designated one of the threads is permitted to issue one or more instructions, but the designated thread that is permitted to issue instructions varies over a plurality of clock cycles in accordance with the instruction issuance sequence. The instructions are pipelined in a manner which permits at least a given one of the threads to support multiple concurrent instruction pipelines.

Abstract: A method and apparatus for efficiently storing multiple data types in a computer's register or data file. A single data file can store data with a variety of sizes and number formats, including integers, fractions, and mixed numbers. The register file is partitioned into fields, such that only the relevant portions of the register file are read or written.

Abstract: A processor having a reduction unit that sums m input operands plus an accumulator value, with the option of saturating after each addition or wrapping around the result of each addition. The reduction unit also allows the m input operands to be subtracted from the accumulator value by simply inverting the bits of the input operands and setting a carry into each of a plurality of reduction adders to one. The reduction unit can be used in conjunction with m parallel multipliers to quickly perform dot products and other vector operations with either saturating or wrap-around arithmetic.

Abstract: A processor comprises a memory, an instruction decoder coupled to the memory for decoding instructions retrieved therefrom, and a plurality of execution units for executing the decoded instructions. One or more of the instructions are in a compound instruction format in which a single instruction comprises multiple operation fields, with one or more of the operation fields each comprising at least an operation code field and a function field. The operation code field and the function field together specify a particular operation to be performed by one or more of the execution units.

Abstract: An arithmetic unit for performing an arithmetic operation on at least first and second input operands, each of the input operands being separable into a first portion and a second portion, such as respective less significant and more significant portions. The arithmetic unit comprises first arithmetic circuitry, second arithmetic circuitry, selection circuitry and saturation circuitry. The first arithmetic circuitry, which may comprise a carry-propagate adder, processes the first portions of the input operands to generate at least a temporary sum and a carry output. The second arithmetic circuitry, which may comprise a dual adder and a preliminary saturation detector, processes the second portions of the input operands to generate one or more temporary sums and a number of saturation flags. The selection circuitry is configured to select one or more of the outputs of the second arithmetic circuitry based on the carry output of the first arithmetic circuitry.

Abstract: An adder circuit includes a plurality of adder stages interconnected in series, with a carry out of each of the adder stages other than a final adder stage being coupled to a carry in of a subsequent one of the adder stages. Carry, generate and propagate signals applied to respective inputs of a carry out computation element in at least a given one of the adder stages are substantially balanced in terms of a number of gate delays experienced by the signals within the adder circuit in arriving at their respective inputs of the carry out computation element. Advantageously, this provides significant reductions in both dynamic switching power and short circuit power in the adder circuit.

Abstract: Techniques for conserving power by controlling program execution in a convergence device comprising a battery or other power source and at least one processor. The processor is configured to perform processing operations associated with voice call communication functions and to perform processing operations associated with data communication functions, and is operative to execute critical programs and noncritical programs. The convergence device stores, for at least a given one of a plurality of noncritical programs associated with the data communication functions, an identifier of at least one alternate capacity program capable of performing substantially the same function as the given program but having a different power source capacity associated therewith.

Abstract: A cache memory for use in a multithreaded processor includes a number of set-associative thread caches, with one or more of the thread caches each implementing a thread-based eviction process that reduces the amount of replacement policy storage required in the cache memory. At least a given one of the thread caches in an illustrative embodiment includes a memory array having multiple sets of memory locations, and a directory for storing tags each corresponding to at least a portion of a particular address of one of the memory locations. The directory has multiple entries each storing multiple ones of the tags, such that if there are n sets of memory locations in the memory array, there are n tags associated with each directory entry. The directory is utilized in implementing a set-associative address mapping between access requests and memory locations of the memory array.

Abstract: A multithreaded processor includes an instruction decoder for decoding retrieved instructions to determine an instruction type for each of the retrieved instructions, an integer unit coupled to the instruction decoder for processing integer type instructions, and a vector unit coupled to the instruction decoder for processing vector type instructions. A reduction unit is preferably associated with the vector unit and receives parallel data elements processed in the vector unit. The reduction unit generates a serial output from the parallel data elements. The processor may be configured to execute at least control code, digital signal processor (DSP) code, Java code and network processing code, and is therefore well-suited for use in a convergence device. The processor is preferably configured to utilize token triggered threading in conjunction with instruction pipelining.

Abstract: Techniques for thread-based memory access by a multithreaded processor are disclosed. The multithreaded processor determines a thread identifier associated with a particular processor thread, and utilizes at least a portion of the thread identifier to select a particular portion of an associated memory to be accessed by the corresponding processor thread. In an illustrative embodiment, a first portion of the thread identifier is utilized to select one of a plurality of multiple-bank memory elements within the memory, and a second portion of the thread identifier is utilized to select one of a plurality of memory banks within the selected one of the multiple-bank memory elements. The first portion may comprise one or more most significant bits of the thread identifier, while the second portion comprises one or more least significant bits of the thread identifier. Advantageously, the invention reduces memory access times and power consumption, while preventing the stalling of any processor threads.

Abstract: A cache memory for use in a multithreaded processor includes a number of set-associative thread caches, with one or more of the thread caches each implementing an eviction process based on access request address that reduces the amount of replacement policy storage required in the cache memory. At least a given one of the thread caches in an illustrative embodiment includes a memory array having multiple sets of memory locations, and a directory for storing tags each corresponding to at least a portion of a particular address of one of the memory locations. The directory has multiple entries each storing multiple ones of the tags, such that if there are n sets of memory locations in the memory array, there are n tags associated with each directory entry. The directory is utilized in implementing a set-associative address mapping between access requests and memory locations of the memory array.

Abstract: Techniques for thread-based register file access by a multithreaded processor are disclosed. The multithreaded processor determines a thread identifier associated with a particular processor thread, and utilizes at least a portion of the thread identifier to select a particular portion of an associated register file to be accessed by the corresponding processor thread. In an illustrative embodiment, the register file is divided into even and odd portions, with a least significant bit or other portion of the thread identifier being used to select either the even or the odd portion for use by a given processor thread. The thread-based register file selection may be utilized in conjunction with token triggered threading and instruction pipelining. Advantageously, the invention reduces register file port requirements and thus processor power consumption, while maintaining desired levels of concurrency.

Abstract: Techniques for conserving power by controlling program execution in a convergence device comprising a battery or other power source and at least one processor. The processor is configured to perform processing operations associated with voice call communication functions and to perform processing operations associated with data communication functions, and is operative to execute critical programs and noncritical programs. The convergence device stores, for at least a given one of a plurality of noncritical programs associated with the data communication functions, an identifier of at least one alternate capacity program capable of performing substantially the same function as the given program but having a different power source capacity associated therewith.

Abstract: The invention provides techniques for reducing the power consumption of pipelined processors. In an illustrative embodiment, the invention evaluates the predicates of predicated instructions in a decode stage of a pipelined processor, and annuls instructions with false predicates before those instructions can be processed by subsequent stages, e.g, by execute and writeback stages. The predicate dependencies can be handled using, e.g., a virtual single-cycle execution technique which locks a predicate register while the register is in use by a given instruction, and then stalls subsequent instructions that depend on a value stored in the register until the register is unlocked. As another example, the predicate dependencies can be handled using a compiler-controlled dynamic dispatch (CCDD) technique, which identifies dependencies associated with a set of instructions during compilation of the instructions in a compiler.

Abstract: Techniques for token triggered multithreading in a multithreaded processor are disclosed. An instruction issuance sequence for a plurality of threads of the multithreaded processor is controlled by associating with each of the threads at least one register which stores a value identifying a next thread to be permitted to issue one or more instructions, and utilizing the stored value to control the instruction issuance sequence. For example, each of a plurality of hardware thread units of the multithreaded processor may include a corresponding local register updatable by that hardware thread unit, with the local register for a given one of the hardware thread units storing a value identifying the next thread to be permitted to issue one or more instructions after the given hardware thread unit has issued one or more instructions. A global register arrangement may also or alternatively be used.

Abstract: Techniques for thread-based register file access by a multithreaded processor are disclosed. The multithreaded processor determines a thread identifier associated with a particular processor thread, and utilizes at least a portion of the thread identifier to select a particular portion of an associated register file to be accessed by the corresponding processor thread. In an illustrative embodiment, the register file is divided into even and odd portions, with a least significant bit or other portion of the thread identifier being used to select either the even or the odd portion for use by a given processor thread. The thread-based register file selection may be utilized in conjunction with token triggered threading and instruction pipelining. Advantageously, the invention reduces register file port requirements and thus processor power consumption, while maintaining desired levels of concurrency.

Abstract: Techniques for thread-based memory access by a multithreaded processor are disclosed. The multithreaded processor determines a thread identifier associated with a particular processor thread, and utilizes at least a portion of the thread identifier to select a particular portion of an associated memory to be accessed by the corresponding processor thread. In an illustrative embodiment, a first portion of the thread identifier is utilized to select one of a plurality of multiple-bank memory elements within the memory, and a second portion of the thread identifier is utilized to select one of a plurality of memory banks within the selected one of the multiple-bank memory elements. The first portion may comprise one or more most significant bits of the thread identifier, while the second portion comprises one or more least significant bits of the thread identifier. Advantageously, the invention reduces memory access times and power consumption, while preventing the stalling of any processor threads.