Motorola Corporation

International Business Machines Corporation

Freescale Semiconductor,inc.

Paul C Rossbach

Terence M Potter

David D Barrera

David S Levitan

Thomas Alan Hoy

Bahador Rastegar

Bradley Gene Burgess

Khoi Mai

Andrew Augustus Bjorksten

Chin-Cheng Kau

Thomas L Thomas, Jr

Jeffrey E Maguire

Jeffrey Pidge Rupley, Ii

Jeffrey T Bridges

Chih-Jiu Peng

Dave Levitan

Bahador Rastegar, Deceased

Albert R Kennedy

Daniel Chen Chow

Chih-Jui Ray Peng

Danny K Jain

A logic circuit including at least one evaluate circuit coupled to a static output logic circuit. In one example, the evaluate circuit includes a dynamic node, a full keeper, an evaluate device, and a logic tree. In some examples, the output logic circuit is a sampled static output logic circuit and includes a sample device. In some examples, the logic circuit includes multiple evaluate circuits, each with a dynamic node coupled to a control gate of a transistor of the output logic circuit. Some examples may include a delay in a clock signal to increase the internal race margin.

Logic circuitry

When a request to branch to an address stored in a return memory location (440) occurs, a busy bit is used to determine whether the return memory location (440) contains updated information. When the information is not updated, a predicted address is provided to the prediction verifier (460) by the link stack (410). Once the busy bit is valid, the prediction verifier (460) determines if a proper prediction was made. When an improper prediction was made, the update portion (415) of the link stack (410) based on information from the comparator (425) determines if a value stored in the link stack (410) matches the value stored in the return memory location (440). The link stack (410) is synchronized based upon a favorable comparison indicating the return memory location value matches a value in the link stack. If a match is not found, the predicted address is placed back on the link stack or alternatively the link stack is cleared.

Data processing system having a synchronizing link stack and method

A method and system in a data processing system of permitting concurrent processing of multiple conditional branch instructions are disclosed. A condition register is established within the processing system. First and second conditional branch instructions are dispatched during a single cycle of the processing system. Prior to speculatively executing the first conditional branch instruction, a first copy of the condition register is stored. Prior to speculatively executing the second conditional branch instruction, a second copy of the condition register is stored. Multiple copies of the condition register are concurrently maintained so that the first and second conditional branch instructions may be concurrently processed during a single cycle of the processing system.

Method and system in data processing system of permitting concurrent

A processor and method for speculatively executing a branch instruction are disclosed. The processor includes a branch prediction unit for predicting a resolution of a speculative branch instruction, which is selectively configurable such that resolution of the speculative branch instruction is predicted in response to only an address of the speculative branch instruction or in response to branch history of at least one previously executed branch instruction. The processor also includes an address calculation unit for determining a target address in response to the predicted resolution of the speculative branch instruction. In one embodiment, the processor further includes configuration logic for dynamically configuring the branch prediction logic.

Processor having a selectively configurable branch prediction unit that

A method and system in a data processing system are disclosed for efficiently handling exceptions. The data processing system includes a register for storing indications of multiple instructions while the multiple instructions are being concurrently processed. An exception is generated within the data processing system. A determination is made whether the exception was generated by one of the multiple instructions. In response to a determination that one of the multiple instructions generated the exception, a determination is then made whether an indication of the instruction which generated the exception is stored in a particular position within a register within the data processing system. In response to a determination that the indication of the instruction is stored in the particular position within the register, the exception is associated with a first priority group. In response to a determination that the indication of the instruction is not stored in the particular position within the register, the exception is associated with a second priority group. In response to a determination that the indication of the instruction did not generate the exception, the exception is associated with the second priority group.

Method and system in a superscalar data processing system for the

A method and system in a data processing system are disclosed for efficiently managing an indication of a status of each of a plurality of registers included with the data processing system. An array is established having multiple entry fields for storing multiple entries. Each of the multiple entry fields is associated with a different one of the plurality of registers. A status of each of the plurality of registers is determined. A plurality of partitions are established within the array. Each of the partitions are concurrently accessible by the data processing system. A plurality of the multiple entry fields are associated with one of the plurality partitions. An entry is stored in each of the multiple entry field. The entry includes the status of each of the plurality of registers. Each entry is associated with one of the partitions so that a plurality of the multiple entries may be concurrently accessed.

Method and system in a data processing system for efficient management

An improved method of addressing within a pipelined processor having an address bit width of m+n bits is disclosed, which includes storing m high order bits corresponding to a first range of addresses, which encompasses a selected plurality of data executing within the pipelined processor. The n low order bits of addresses associated with each of the selected plurality of data are also stored. After determining the address of a subsequent datum to be executed within the processor, the subsequent datum is fetched. In response to fetching a subsequent datum having an address outside of the first range of addresses, a status register is set to a first of two states to indicate that an update to the first address register is required. In response to the status register being set to the second of the two states, the subsequent datum is dispatched for execution within the pipelined processor. The n low order bits of the subsequent datum are then stored, such that memory required to store addresses of instructions executing within the pipelined processor is thereby decreased.

Method and system of addressing which minimize memory utilized to store

A cache memory uses content-addressable tag-compare arrays (CAM) to determine if a match occurs. In a semi-associative instruction cache, with the CAM and eight cache lines grouped together to form camlets, a binary index is used to address one camlet in the cache array, and the effective address tag match is used to select a potential line within the camlet in accessing data stored in the cache array. Since an E-tag match causes that cache line's wordline to activate, proper cache operation requires that no two (or more) E-tags within a camlet have the same match criteria (ECAM entry); the invalidation of entries is done to prevent this from happening. Due to the mapping of the effective address into the E-tag CAM and the camlet binary index, addresses that are 1-Meg apart point to the same camlet and have the same ECAM tag. The method thus employs a semi-associative cache having cache lines configured in camlets of, for example, eight lines per camlet. An LRU indication is stored in each camlet showing which line was least-recently-used. Upon occurrence of a cache replacement operation, it is determined whether or not a replacement line has a tag matching a line that is already in the camlet, and, if so, this line is invalidated and it is indicated to be the least-recently-used line. The next replacement goes to this line, whereas otherwise it would have appeared to be the most-recently-used since its wordline went high for the invalidate operation.

Semi-associative cache with MRU/LRU replacement

A data processor (12) has a branch prediction unit (28) that predicts conditional branch instructions and a control unit (70) therein that monitors the number of unresolved branch instructions. This control unit selectively allows the data processor to fetch the instructions indicated by the branch prediction unit from an external memory system depending upon the number of unresolved branch instructions. The particular threshold number of unresolved branch instructions is user programmable. The data processor thereby limits its bus accesses to those occasions when it is reasonably sure that it will need the indicated instructions.

Data processor with programmable levels of speculative instruction

A memory cache (14) has a semi-associative cache array (50), a cache reload buffer (40), and a cache reload buffer driver (42). The memory cache writes received data to the cache reload buffer and waits until the data is requested again before it invalidates any cache aliased entries in the semi-associative cache array. This invalidation step requires no dedicated cycle but instead is a result of the memory cache being able to simultaneously read from the semi-associative cache array and the cache reload buffer.

Memory cache with automatic alliased entry invalidation and method of

A CAM/SRAM structure (42) performs address translations of variable length blocks, a 'block address translator.' Each address translation is stored in a register broken into an upper half and a lower half. The upper half contains CAM bit cells (56) which match an input effective address to a stored tag (BEPI) alternating with SRAM bit cells which store a block length tag (BL). The block length tag defines the length of the translated block and, hence, the number of bits which must match between the input effective address and the stored tag. The lower half contains SRAM bit cells which store a real address associated with the tag (BRPN) alternating with multiplexer circuits. In the event of a CAM match, each multiplexer circuit outputs either a real address bit or an input effective address bit, depending upon the block length tag. The two halves of each register are fabricated adjacent to each other in parallel rows to minimize routing requirements and reduce overall circuit capacitance.

Address translation circuit

A memory cache (14) has a plurality of cache lines (50) for storing a series of contiguous memory elements. Each series of memory elements are interlaced within the corresponding cache line on a element-by-element basis and on a bit-by-bit basis. This storage strategy allows the memory cache to output a subset memory elements within a cache line quickly and in the original contiguous order. The invention may be advantageously incorporated in an instruction cache of superscalar data processor to provide a series of sequential instructions for execution.

Memory cache with interlaced data and method of operation

A precharge device (28) has a first (30) and a second node (32), a transistor tree (29), a screening transistor (Q20) and clocking circuitry (Q17, Q18, Q19). The transistor tree (29) couples the first (30) and the second (32) node and is operable to electrically short-circuit the nodes according to input signals (A.sub.1, A.sub.2, A.sub.3). The screening transistor (Q20) has a first and a second [source-drain region] current electrode and a [gate] control electrode. The first [source-drain region] current electrode is coupled to a third node (34), the second [source-drain region] current electrode is coupled to the second node (32) and the [gate] control electrode is coupled to the first node (30). The clocking circuitry alternately precharges the first (30) and third nodes (34) to a first known voltage level and evaluates the voltage on the first node (30) to output a logic level.

Decoder/comparator and method of operation

A method and apparatus for performing integer and floating-point divide operations using a single modified SRT divider in a data processor. The floating-point and integer division is performed using SRT division on normalized positive mantissas (dividend and divisor). Integer division shares portions of the floating point circuitry, however, the sequence of operations is modified during the performance of an integer divide operation. The SRT divider performs a sequence of operations before and after an iteration loop to re-configure an integer divisor and dividend into a data path representation which the SRT algorithm requires for floating-point mantissas. During the iteration loop, quotient bits are selected and used to generate intermediate partial remainders. The quotient bits are also input to quotient registers which accumulate the final quotient mantissa. A full mantissa adder is used to generate a final remainder.

Method and apparatus for performing integer and floating point division

In a data processor an Sweeney-Robertson-Tocher (SRT) divider is provided having a negative divisor sticky detection circuit. The negative divisor sticky detection circuit allows negative sticky correction to occur in the SRT divider without requiring additional iteration cycles. At the conclusion of iterative cycles of a divide operation, a final remainder is formed and stored in a latch in the SRT divider. The negative divisor sticky detection circuit determines whether a negative final remainder is equal in magnitude to a divisor value by bit-wise XORing the final remainder with the two's complement value of a divisor, immediately before sticky logic detects the negative sticky bit. The final sticky value is obtained by logically combining the negative sticky bit with a positive sticky bit computed by a positive divisor sticky detection circuit.

Growing community of inventors

Austin, TX, United States of America

Paul C Rossbach