Abstract: The decoding apparatus in the present invention includes a memory operable to hold encoded data representing one of a compressed sound and a compressed image, a memory read-out unit operable to sequentially read out the encoded data from said memory, a match determining circuit operable to determine whether or not data matching a specific bit sequence exists in the encoded data read out by said memory read-out unit, a deleting circuit operable to delete a part of the specific bit sequence from the encoded data read out from said memory, when said match determining circuit determines that the specific bit sequence exists, and a decoding circuit operable to decode the post-deletion encoded data.

Abstract: A system and method for extracting complex, variable length computer instructions from a stream of complex instructions each subdivided into a variable number of instructions bytes, and aligning instruction bytes of individual ones of the complex instructions. The system receives a portion of the stream of complex instructions and extracts a first set of instruction bytes starting with the first instruction bytes, using an extract shifter. The set of instruction bytes are then passed to an align latch where they are aligned and output to a next instruction detector. The next instruction detector determines the end of the first instruction based on said set of instruction bytes. An extract shifter is used to extract and provide the next set of instruction bytes to an align shifter which aligns and outputs the next instruction. The process is then repeated for the remaining instruction bytes in the stream of complex instructions.

Abstract: A new technique for accelerating the computational speed of a computer algorithm is provided. The inventive technique can be applied to an encryption algorithm, such as an RSA algorithm, for example. A method and system for effecting secure information communications utilizing an accelerated information encryption algorithm is also contemplated.

Abstract: A data processor according to the present invention executes instructions described in first and second instruction formats. The first instruction format defines a register-addressing field of a predetermined size, while the second instruction format defines a register-addressing field of a size larger than that of the register-addressing field defined by the first instruction format. The data processor includes: instruction-type identifier, responsive to an instruction, for identifying the received instruction as being described in the first or second instruction format by the instruction itself; a first register file including a plurality of registers; and a second register file also including a plurality of registers, the number of the registers included in the second register file being larger than that of the registers included in the first register file.

Abstract: An apparatus and method are provided for extending a microprocessor instruction set to allow for selective suppression of store checking at the instruction level. The apparatus includes fetch logic, and translation logic. The fetch logic receives an extended instruction. The extended instruction has an extended prefix and an extended prefix tag. The extended prefix specifies that store checking be suppressed for the extended instruction. The extended prefix tag is an otherwise architectural opcode within an existing instruction set. The fetch logic precludes store checking for pending store events associated with the extended instruction. The translation logic is coupled to the fetch logic. The translation logic translates the extended instruction into a micro instruction sequence that sequence directs the microprocessor to exclude store checking during execution of a prescribed operation.

Abstract: An apparatus and method are provided for extending a microprocessor instruction set to specify non-temporal memory references at the instruction level. The apparatus includes translation logic and extended execution logic. The translation logic translates an extended instruction into a micro instruction sequence. The extended instruction has an extended prefix and an extended prefix tag. The extended prefix specifies a non-temporal access for a memory reference prescribed by the extended instruction, where the non-temporal access cannot be specified by an existing instruction from an existing instruction set. The extended prefix tag indicates the extended prefix, where the extended prefix tag is an otherwise architecturally specified opcode within the existing instruction set. The extended execution logic is coupled to the translation logic. The extended execution logic receives the micro instruction sequence, and executes the non-temporal access to perform the memory reference.

Abstract: Methods and apparatus relating to speculatively decoding instruction lengths in order to increase instruction throughput are described. In an embodiment, instructions are speculatively decoded within a pipelined microprocessor architecture such that up to four instruction lengths may be decoded within a maximum of two processor clock cycles. Other embodiments are also disclosed.

Abstract: An embodiment of the invention relates to a process for compression of executable code by a microprocessor, comprising decomposing the executable code into words; dividing the executable code into instruction lines; compressing each word of each line in the form of a compressed word of variable length, the compressed words of a line being combined into a line of compressed words; and constituting an addressing table localizing each of the lines of compressed words in a block of lines compressed words and comprising one input per group of lines of compressed words, each input (j) specifying the position of a first line of compressed words in the block, and the respective lengths of the lines of compressed words of group, except for a last line of compressed words of the group, whereof the length is determined by means of the position of a first line of compressed words of a following group.

Abstract: When data is transferred to an access destination in a different endian format, a transfer start address is aligned based on a transfer bus width, and a transfer size is adjusted according to the transfer bus width and a transfer address. Thus, it becomes possible to perform burst transfer in the access destination. Accordingly, in the case where burst transfer to an access destination in a different endian format is performed with a smaller data width than a transfer bus width, an inconvenience where burst transfer can not be performed because an address is converted and data access is no longer an ascending order access can be prevented.

Abstract: As shown in FIG. 1, an operation-processing device of the present invention comprises a register array (11) having plural registers for holding an arbitrary value based on a write address Aw and a write control signal Sw and outputting this value based on a read address Ar, an ALU (12) for performing operations on this value, a decoder (13) for decoding an operation instruction from an operation program AP for operating this ALU (12), and an instruction-execution-controlling portion (50) for controlling the register array (11) and the ALU (12) in order to execute this operation instruction, wherein this instruction-execution-controlling portion (50) selects one of the registers based on the operation instruction and performs register-to-register addressing processing that, based on a value held by this selected register, selects another register.

Abstract: A method and apparatus for loading and storing vectors from and to memory, including embedding a location identifier in bits comprising a vector load and store instruction, wherein the location identifier indicates a location in the vector where useful data ends. The vector load instruction further includes a value field that indicates a particular constant for use by the load/store unit to set locations in the vector register beyond the useful data with the constant. By embedding the ending location of the useful date in the instruction, bandwidth and memory are saved by only requiring that the useful data in the vector be loaded and stored.

Abstract: A load/store unit includes a Top register for storing a value retained before loading to a load destination register and a saved register capable of storing data retained to the Top register. When an unaligned instruction evaluation unit determines that a load instruction issued from a instruction decode unit is an unaligned instruction, data stored to the Top register are stored to the saved register in order to make the Top register available to subsequent load instructions issued from the instruction decode unit.

Abstract: An apparatus and method for extending a microprocessor instruction set is provided. The apparatus includes translation logic and extended execution logic. The translation logic translates an extended instruction into corresponding micro instructions. The extended instruction has an extended prefix and an extended instruction tag. The extended prefix directs that an architectural extension be employed in the execution of an operation prescribed by the extended instruction. The extended instruction tag indicates the extended instruction prefix, where the extended instruction tag is an otherwise architecturally specified opcode within the microprocessor instruction set. The extended execution logic is coupled to the translation logic, and receives the corresponding micro instructions, and employs the architectural extension in the execution of the operation.

Abstract: Apparatus and method are provided for extending a microprocessor instruction set to allow for instruction-level specification of floating point format to be employed during execution of an associated floating point operation. The apparatus includes translation logic and extended execution logic. The translation logic translates an extended instruction into corresponding micro instructions. The extended instruction has an extended prefix and an extended prefix tag. The extended prefix specifies the floating point format. The extended prefix tag indicates the extended prefix, where the extended prefix tag is an otherwise architecturally specified opcode within an instruction set for a microprocessor. The extended execution logic is coupled to the translation logic. The extended execution logic receives the corresponding micro instructions, and executes the associated floating point operation according to the floating point format specified by the extended prefix.

Abstract: Techniques for processing transmissions in a communications (e.g., CDMA) system. The method and system encode and process instructions of mixed lengths (e.g., 16 bits and 32 bits) and instruction packets including instructions of mixed lengths. This includes encoding a plurality of instructions of a first length and a plurality of instructions of a second length. The method and system encode a header having at least one instruction length bit. The instruction bit distinguishes between instructions of the first length and instructions of the second length for an associated DSP to process in a mixed stream. The method and system distinguish between the instructions of the first length and the instructions of the second length according to the contents of the instruction length bits. The header further includes bits for distinguishing between instructions of varying lengths in an instruction packet.

Abstract: The present invention includes a method and device for controlling the data length of read and write operations performed on a memory device. The method includes determining a first number of channels available to a memory controller operatively coupled to the memory device; determining a second number representative of the number of populated channels; calculating a burst length based on the first and second numbers; and programming the memory controller to use the burst length as the data length of read and write operations performed on the memory device.

Abstract: The present invention discloses a method for compressing instruction codes. This method comprises: compressing an instruction block including a plurality of instructions according to Huffman-Encoding technique; determining whether it's necessary to insert no-operation (nop) instructions among the plurality of compressed instructions according to a compression ratio, so as to generate a plurality of new instruction blocks complying with the compression ratio; if it's necessary to insert nop instructions, inserting nop instructions among the plurality of compressed instructions to form the plurality of new instruction blocks; and repeating the above-mentioned steps until no nop instructions have to be inserted.

Abstract: A digital signal processor uses a variable length instruction set. The variable length instructions may be stored in adjacent locations within memory space. The beginning and ending of instructions may, but are not required to, occur across memory word boundaries. Preferably, the variable length instructions contain variable numbers of instruction fragments. Each instruction fragment causes a particular operation (or operations) to be performed, thereby allowing multiple operations to be performed during each clock cycle. This reduces the total number of clock cycles necessary to perform a task.

Abstract: An instruction length determination device includes an instruction input unit having a memory space to store a plurality of N-bit data; an instruction fetch unit which fetches the plurality of N-bit data from the instruction input unit; an instruction length determination logic which compares concatenate bits of a first N-bit data with a predetermined value for determination of an instruction length; and an instruction concatenate unit which selectively concatenates a number of successive N-bit data based on the determination. The instruction length determination logic determines that the first N-bit data is a complete instruction when the concatenate bit of the first N-bit data is not equal to the predetermined value. Otherwise, the instruction length determination logic determines that a complete instruction is formed of last N-bit data finally fetched and all N-bit previously reserved.

Abstract: Apparatus and methods for real-time control using a data processor. In one aspect, the invention comprises an improved processor having one or more extension instructions (and associated supporting pipeline hardware) which are specially adapted for use in a real-time control algorithm running on the processor. In one exemplary embodiment, the processor is a 32-bit pipelined RISC device having custom multiply (CMUL) and multiply-accumulate (CMAC) instructions added to the extension instruction set to optimize algorithm performance in real-time linear time-invariant (LTI) applications. Specialized extension hardware, and methods for generating a processor design adapted for real-time control applications are also disclosed.

Abstract: A processor 2 is provided with the ability to execute program instructions in the form of Java bytecodes including a dedicated null checking instruction. The null checking instruction reads the top of stack value, compares this with a null value and jumps to an exception handling routine if the top of stack value equals the null value, otherwise the next program instruction is executed.

Abstract: A digital processor and method of operation utilize an alias address space to implement variable length instruction encoding on a legacy processor. The method includes storing instructions of a code sequence in memory; generating instruction addresses of the code sequence; automatically switching between a first operating mode and a second operating mode in response to a transition in instruction addresses between a first address space and a second address space, wherein addresses in the first and second address spaces access a common memory space; in the first operating mode, accessing instructions in the first address space; in the second operating mode, accessing instructions in the second address space; and executing the accessed instructions of the code sequence. Instructions of different instruction lengths may be utilized in the first and second operating modes.

Abstract: Expanding the capacity of a fixed digital field using a unique number calculated from the digital field, such as an error code. Expansion is possible by calculating a new error code using a different algorithm. A recipient, upon not detecting the error code from the original algorithm, checks for the new error code before indicating an error. The presence of the new error code acts like an extra bit to give an entirely new set of numbers for the digital field, thus doubling the command set. In another aspect of the invention, fill bits between transmission packets are used to indicate further data, and are included in calculating the new error code.

Abstract: A mechanism for superscalar decode of variable length instructions. The decode mechanism may be included within a processing unit, and may comprise a length decode unit. The length decode unit may obtain a plurality of instruction bytes. The instruction bytes may be associated with a plurality of variable length instructions, which are to be executed by the processing unit. The length decode unit may perform a length decode operation for each of the plurality of instruction bytes. For each instruction byte, the length decode unit may estimate the instruction length of a current variable length instruction associated with a current instruction byte. Furthermore, during the length decode operation, for each instruction byte, the length decode unit may estimate the start of a next variable length instruction based on the estimated instruction length of the current variable length instruction, and store a first pointer to the estimated start of the next variable length instruction.

Abstract: A method, system, and computer program product are provided for identifying instructions to obtain representative traces. A phase instruction budget is calculated for each phase in a set of phases. The phase instruction budget is based on a weight associated with each phase and a global instruction budget. A starting index and an ending index are identified for instructions within a set of intervals in each phase in order to meet the phase instruction budget for that phase, thereby forming a set of interval indices. A determination is made as to whether the instructions within the set of interval indices meet the global instruction budget. Responsive to the global instruction budget being met, the set of interval indices are output as collection points for the representative traces.

Abstract: In one embodiment, a rotator, a mask generator, and circuitry configured to mask the rotated operand output by the rotator with the output mask generated by the mask generator perform a shift operation. Coupled to receive the input operand and the shift count, the rotator is configured to rotate the input operand by the shift count. Coupled to receive the shift count and the shift direction, the mask generator is configured to generate an output mask by decoding a most significant bit (MSB) field of the shift count to generate a first mask, decoding a least significant bit (LSB) field of the shift count to generate a second mask, logically ANDing the bits of the second mask with the corresponding bit of the first mask and logically ORing the result with an adjacent bit of the first mask that is selected responsive to the shift direction.

Abstract: A processor and its arithmetic instruction processing method and arithmetic operation control method are disclosed that add a new operand designation option to SIMD arithmetic instructions and permit software pipelining between arithmetic operations performed in parallel by a SIMD arithmetic unit. A selector for adding an operation for interchanging multiple outputs of a SIMD arithmetic unit is added to a data path. A register file is divided in accordance with the output bit fields of the SIMD arithmetic unit. A means of specifying multiple registers as a SIMD instruction's output operand is added. Therefore, part of the output results of arithmetic operations performed in parallel by the SIMD arithmetic unit can be stored in a register providing the input for another arithmetic operation. Software pipelining is rendered achievable in this manner.

Abstract: A data processing apparatus 2 is provided which is responsive to data access instructions to perform data access operations. These data access instructions have a first form utilizing a 12-bit offset field but with a fixed addressing mode and a second form utilizing a shorter 8-bit offset field but with an addressing mode specified within a manipulation mode control field of the data access instruction.

Abstract: In a processor executing instructions from a variable-length instruction set, a preload instruction is operative to retrieve from memory a data block corresponding to an instruction cache line, pre-decode instructions from a variable-length instruction set in the data block, and load the instructions and pre-decode information into the instruction cache. An instruction execution unit indicates to a pre-decoder the position within the data block of a first valid instruction. The pre-decoder successively determines the length of each instruction and hence the instruction boundaries. An instruction cache line offset indicator that identifies the position of the first valid instruction may be generated and provided to the pre-decoder in a variety of ways.

Abstract: A set of low-cost microcontroller extensions facilitates Digital Signal Processing (DSP) applications by incorporating a Multiply-Accumulate (MAC) unit in a Central Processing Unit (CPU) of the microcontroller which is responsive to the extensions.

Abstract: Parallel table lookups are implemented using variable Mux instructions to reorder data. Table data can be represented in a “table” register, while the desired ordering can be represented in an “Index” register. A direct variable Mux instruction can specify the table register and the index register as arguments, along with a result register. The instruction writes at least some of the data from the table register into the result register as specified in the index register. If the entire table cannot fit within a single register, entries can be divided between two or more table registers. An indirect variable Mux instruction can specify both a table-register-select register and a subword-location-select register. Both the direct and indirect Mux instructions can be used with entry data that is divided in accordance with significance between registers. In that case, plural Mux instructions are used with UnPack instructions that concatenate portions of the table entries.

Abstract: There are provided methods and computer program products for implementing instruction set architectures with non-contiguous register file specifiers. A method for processing instruction code includes processing a fixed-width instruction of a fixed-width instruction set using a non-contiguous register specifier of a non-contiguous register specification. The fixed-width instruction includes the non-contiguous register specifier.

Abstract: Systems and apparatuses are presented relating a programmable processor comprising an execution unit that is operable to decode and execute instructions received from an instruction path and partition data stored in registers in the register file into multiple data elements, the execution unit capable of executing a plurality of different group floating-point and group integer arithmetic operations that each arithmetically operates on multiple data elements stored registers in a register file to produce a catenated result that is returned to a register in the register file, wherein the catenated result comprises a plurality of individual results, wherein the execution unit is capable of executing group data handling operations that re-arrange data elements in different ways in response to data handling instructions.

Abstract: A microprocessor apparatus and method are provided, for selectively controlling write back of a result. The apparatus includes translation logic and extended execution logic. The translation logic translates an extended instruction into corresponding micro instructions. The extended instruction has an extended prefix and an extended prefix tag. The extended prefix precludes write back of the result, where the result is that which is produced by executing an operation prescribed by said extended instruction, and wherein the result would otherwise be written back into a destination register. The extended prefix tag indicates the extended prefix, where the extended prefix tag is an otherwise architecturally specified opcode within an instruction set for a microprocessor. The extended execution logic is coupled to the translation logic. The extended execution logic receives the corresponding micro instructions, and executes the operation to generate the result, and precludes write back of the result.

Abstract: Variable length instructions are formed for execution in a processing system. Each instruction includes a parameter portion having one or more of predetermined types of parameters and an opcode portion. The opcode portion specifies an operation to be performed, the number of parameters in the instruction, and definitive characteristics of the parameters. The parameters may represent data which is compressible, thereby enabling the size of parameters in an instruction to be reduced.

Abstract: An apparatus and method are provided, for accessing extended registers within a microprocessor. The apparatus includes translation logic and extended register logic. The translation logic translates an extended instruction into corresponding micro instructions for execution by the microprocessor. The extended instruction has an extended prefix and an extended prefix tag. The extended prefix specifies register address extensions, the register address extensions indicating the extended registers, where the extended registers cannot be specified by an existing instruction set, and where the existing instruction set includes the x86 instruction set. The extended prefix tag indicates the extended prefix, where the extended prefix tag is an otherwise architecturally specified opcode within the existing instruction set, and where the extended prefix tag includes opcode F1 (ICE BKPT) in the x86 instruction set. The extended register logic is coupled to the translation logic.

Abstract: In one embodiment, a state machine receives a plurality of instructions from an instruction register to be processed by a digital signal processor. After receiving a single RTI, the state machine loads each of the plurality of instructions one at time and determines the validity of each instruction. If the instruction is valid, the state machine transfers the instruction to the decoder. If the instruction is invalid or if a no-operation instruction is present, the state machine discards the instruction and immediately loads the next instruction.

Abstract: A data processing system including an instruction cache 8 and an instruction decompression circuit 10 between the instruction cache 8 and a compressed instruction data memory 12. The instruction decompression circuit decompresses compressed instruction data CID recovered from the compressed instruction data memory and forms program instructions which are supplied to the instruction cache. The program instructions are compressed in blocks of program instructions with an associated mask value where the bit values within the mask indicate whether corresponding bit slices within the blocks of program instructions are to be represented by a default bit value or a separately specified by bit slice specifier values. This technique is particularly well suited to VLIW processors.

Abstract: A method comprising fetching an input from at least one of a plurality of floating-point registers and detecting whether the input includes a token. If the token is detected in the input, checking what mode the processor is in. If the processor is in a first mode, processing the input to render an arithmetic result. If the processor is in a second mode, performing a token specific operation. And producing an output. The present invention also provides a processor comprising a first instruction set engine, a second instruction set engine, and a mode identifier. A plurality of floating-point registers are shared by the first instruction set engine and the second instruction set engine. A floating-point unit is coupled to the floating-point registers. The floating-point unit processes an input responsive to the mode identifier and the input to produce an output.

Abstract: Method, apparatus, and program means for performing a string comparison operation. In one embodiment, an apparatus includes execution resources to execute a first instruction. In response to the first instruction, said execution resources store a result of a comparison between each data element of a first and second operand corresponding to a first and second text string, respectively.

Abstract: A method for processing a variable length code comprising: determining a first address; decoding opcodes from the at least one table starting at a first address; in response to each of the opcodes: receiving a portion of a sequence of bits, the sequence of bits comprising a first variable length code; receiving S from the second table at the current address; flushing S bits in the sequence of bits; receiving T corresponding to one of the stages; determining a value of a set of T bits in the sequence of bits; receiving D from the second table at the current address; and computing the next address, the next address being the sum of the current address, D, and the value of the set of T bits; and retrieving the next opcode, the next opcode being retrieved from the next address; and determining the decoded syntax element.

Abstract: A system and method for extracting complex, variable length computer instructions from a stream of complex instructions each subdivided into a variable number of instructions bytes, and aligning instruction bytes of individual ones of the complex instructions. The system receives a portion of the stream of complex instructions and extracts a first set of instruction bytes starting with the first instruction bytes, using an extract shifter. The set of instruction bytes are then passed to an align latch where they are aligned and output to a next instruction detector. The next instruction detector determines the end of the first instruction based on said set of instruction bytes. An extract shifter is used to extract and provide the next set of instruction bytes to an align shifter which aligns and outputs the next instruction. The process is then repeated for the remaining instruction bytes in the stream of complex instructions.

Abstract: Instructions of a program are stored in compressed form in a program memory (12). In a processor which executes the instructions, a program counter (50) identifies a position in the program memory. An instruction cache (40) has cache blocks, each for storing one or more instructions of the program in decompressed form. A cache loading unit (42) includes a decompression section (44) and performs a cache loading operation in which one or more compressed-form instructions are read from the position in the program memory identified by the program counter and are decompressed and stored in one of the said cache blocks of the instruction cache. A cache pointer (52) identifies a position in the instruction cache of an instruction to be fetched for execution. An instruction fetching unit (46) fetches an instruction to be executed from the position identified by the cache pointer.

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: The present application describes a method and a processor for handling register dependency conflicts between lesser and greater width instructions, colloquially referred to as “evil twins.” If there is a register dependency between a greater width producer instruction and a lesser width consumer instruction, a greater width source register is substituted for the source register specified by the lesser width producer. If there is a register dependency between a lesser width producer instruction and a greater width producer instruction, the greater width consumer instruction is replaced by multiple helper instructions. One or more of the helper instructions merge lesser width registers aliased onto the source registers specified by the greater width consumer instruction, into temporary registers. Another helper instruction executes the greater width consumer instruction using the temporary registers instead of the original source registers.

Abstract: A data processing device has an instruction decoder (1), a control logic unit (3), and ALU (4). The instruction decoder (1) decodes instruction codes of an arithmetic instruction. The control logic unit (3) detects the effective data width of operation data to be processed according to the decode result from the instruction decoder (1) and determines the number of cycles for the instruction execution corresponding to the effective data width. The ALU (4) executes the instruction with the number of cycles of the instruction execution determined by the control logic unit (3).

Abstract: An apparatus and method are provided for extending a microprocessor instruction set to specify non-temporal memory references at the instruction level. The apparatus includes translation logic and extended execution logic. The translation logic translates an extended instruction into a micro instruction sequence. The extended instruction has an extended prefix and an extended prefix tag. The extended prefix specifies a non-temporal access for a memory reference prescribed by the extended instruction, where the non-temporal access cannot be specified by an existing instruction from an existing instruction set. The extended prefix tag indicates the extended prefix, where the extended prefix tag is an otherwise architecturally specified opcode within the existing instruction set. The extended execution logic is coupled to the translation logic. The extended execution logic receives the micro instruction sequence, and executes the non-temporal access to perform the memory reference.

Abstract: Speculatively decoding instruction lengths in order to increase instruction throughput. Instructions are speculatively decoded within a pipelined microprocessor architecture such that up to four instruction lengths may be decoded within a maximum of two processor clock cycles.

Abstract: The present invention discloses a program converting unit for generating a machine language instruction from a source program for a processor that manages an N-bit address while processing M-bit data, N being greater than M, and such a processor that runs the converted program. The program converting unit comprising: a parameter holding unit for holding a data width and a pointer width designated by a user; the data width representing the number of bits of data used in the source program while the pointer width representing the number of bits of an address; and a generating unit for generating an instruction to manage the data width when a variable operated by the instruction represents the data, and for generating an instruction to manage the pointer width when a variable operated by the instruction represents the address.

Abstract: A reconfigurable register file integrated in an instruction set architecture capable of extended precision operations, and also capable of parallel operation on lower precision data is described. A register file is composed of two separate files with each half containing half as many registers as the original. The halves are designated even or odd by virtue of the register addresses which they contain. Single width and double width operands are optimally supported without increasing the register file size and without increasing the number of register file ports. Separate extended registers are also employed to provide extended precision for operations such as multiply-accumulate operations.