Intel Corporation

Sunil R Shenoy

Ashwani Kumar Gupta

David B Papworth

Michael A Fetterman

Reynold V D'Sa

Bradley D Hoyt

Subramanian Natarajan

Glenn J Hinton

Jay S Heeb

Scott B Huck

Mark S Myers

Jimmy W Wong

M Vittal Kini

Glenn I Hinton

Konrad K Lai

Robert C Duzett

Frank S Smith

Larry O Smith

Atiq A Bajwa

Kent Mason

James W Wong

A Branch Target Buffer Circuit in a computer processor that predicts branch instructions with a stream of computer instructions is disclosed. The Branch Target Buffer Circuit uses a Branch Target Buffer Cache that stores branch information about previously executed branch instructions. The branch information stored in the Branch Target Buffer Cache is addressed by the last byte of each branch instruction. When an Instruction Fetch Unit in the computer processor fetches a block of instructions it sends the Branch Target Buffer Circuit an instruction pointer. Based on the instruction pointer, the Branch Target Buffer Circuit looks in the Branch Target Buffer Cache to see if any of the instructions in the block being fetched is a branch instruction. When the Branch Target Buffer Circuit finds an upcoming branch instruction in the Branch Target Buffer Cache, the Branch Target Buffer Circuit informs an Instruction Fetch Unit about the upcoming branch instruction.

Method and apparatus for implementing a set-associative branch target

Method and apparatus for implementing a branch target buffer in CISC

Streaming buffer renaming for memory accesses issued by a microprocessor to an external memory via a system bus allows up to N fetch accesses at any one time for M physical streaming buffer locations, where N is greater than M. When a fetch within the processor misses the instruction cache, the fetch address is placed in the streaming buffer. When the data has been fetched from the external memory, it is returned to the streaming buffer and placed into one of the M physical buffer locations. The data within the streaming buffer is returned to the instruction cache of the processor only if it is to be used in accordance with the computer program being executed.

Computer system employing streaming buffer for instruction preetching

A four stage branch instruction resolution system for a pipelined processor is disclosed. A first stage of the branch instruction resolution system predicts the existence and outcome of branch instructions within an instruction stream such that an instruction fetch unit can continually fetch instructions. A second stage decodes all the instructions fetched. If the decode stage determines that a branch instruction predicted by the first stage is not a branch instruction, the decode stage flushes the pipeline and restarts the processor at a corrected address. The decode stage verifies all branch predictions made by the branch prediction stage. Finally, the decode stage makes branch predictions for branches not predicted by the branch prediction stage. A third stage executes all the branch instructions to determine a final branch outcome and a final branch target address. The branch execution stage compares the final branch outcome and final branch target address with the predicted branch outcome and predicted branch target address to determine if the processor must flush the front-end of the microprocessor pipeline and restart at a corrected address. A final branch resolution stage retires all branch instructions. The retirement stage ensures that any instructions fetched after a mispredicted branch are not committed into permanent state.

Dual prediction branch system having two step of branch recovery process

A method and apparatus for resolving Return From Subroutine instructions in a computer processor are disclosed. The method and apparatus resolve Return From Subroutine instructions in four stages. A first stage predicts Call Subroutine instructions and Return From Subroutine instructions within the instruction stream. The first stage stores a return address in a return register when a Call Subroutine instruction is predicted. The first stage predicts a return to the return address in the return register when a Return From Subroutine instruction is predicted. A second stage decodes each Call Subroutine and Return From Subroutine instruction in order to maintain a Return Stack Buffer that stores a stack of return addresses. Each time the second stage decodes a Call Subroutine instruction, a return address is pushed onto the Return Stack Buffer. Correspondingly, each time the second stage decodes a Return From Subroutine instruction, a return address is popped off of the Return Stack Buffer. The second stage verifies predictions made by the first stage and predicts return addresses for Return From Subroutine instructions that were not predicted by the first stage. A third stage executes Return From Subroutine instructions such that the predictions are verified. Finally, a fourth stage retires Return From Subroutine instructions and ensures that no instructions fetch after a mispredicted return address are committed into permanent state.

Method and apparatus for predicting and handling resolving return from

A microprocessor and method which allows data consistency to be maintained between a memory which is external to the microprocessor and a data cache unit. The microprocessor has a central processing unit coupled to a local bus. A direct memory access unit coupled to the central processing unit for loading data from and storing data to the direct access memory unit. The local bus is coupled to a system bus and has a bus control unit controlling the loading and storing of data on the system bus. The system bus transfers data external to the microprocessor using the bus control unit upon instructions from the central processing unit. A data cache unit is coupled to the local bus and selectively stores a copy of data loaded by the bus control unit and receives a memory address from the local bus during a memory access by either the central processing unit or the direct memory access unit. The microprocessor employs a mechanism that invalidates copy data when the memory access is a store by the direct memory access unit when a cache hit is detected. Further, the microprocessor employs a mechanism that designates as non-cacheable the loading of data by the direct access memory unit, even though the data was previously designated as cachable, preventing the data cache unit from performing any action, namely the overwriting of more critical data within the data cache unit.

Method and apparatus for using a direct memory access unit and a data

Method and apparatus for resolving return from subroutine instructions

A dynamically expandable pipeline in a microprocessor. The present invention is used in a microprocessor or a microprocessor in a computer system. The present invention delays execution of a cacheable LOAD instruction by a bus controller for one cycle to allow sufficient time for 'hit or miss' detection by a data cache unit. The present invention dynamically expands the instruction pipeline for cacheable LOAD instructions that 'miss' an on-chip data cache when the LOAD is followed by another instruction that uses the bus controller. The dynamic pipeline allows time for the 'hit or miss' detection by the data cache unit without unnecessarily degrading pipeline performance. The present invention offers increased overall microprocessor and computer system performance by allowing efficient implementation of an on-chip data cache. The present invention provides increased performance without undue or overly complex modifications to existing pipeline or data cache circuits.

Method and apparatus for dynamically expanding the pipeline of a

A method and apparatus for performing bi-endian byte and short accesses in a single endian microprocessor. The present invention is used in a microprocessor or in a microprocessor in a computer system. The present invention provides a single endian microprocessor that promotes sub-word accesses to word accesses with a means for manipulating the two least significant bits of the access address to point to the correct sub-word data returned during an access to bi-endian external memory. The method for manipulating the address bits is also used to allow a single endian data cache to operate with the bi-endian external memory. The two LSBs of the address are manipulated such that the pointer values are A1# and A0# for word promoted byte accesses or cacheable accesses. For word promoted short accesses or cacheable accesses, the pointer values are A1# and A0. The present invention offers increased flexibility in interfacing a single-endian microprocessor with bi-endian systems. The present invention provides easy interfacing without undue or overly complex modifications to existing circuits.

Method and apparatus for performing bi-endian byte and short accesses in

An instruction locking apparatus and method for a cache memory allowing execution time predictability and high speed performance. The present invention implements a cache locking scheme in a two set associative instruction cache that utilizes a specially designed Least Recently Used (LRU) unit to effectively lock a first portion of the instruction cache to allow high speed and predictable execution time for time critical program code sections residing in the first portion while leaving another portion of the instruction cache free to operate as an instruction cache for other, non-critical, code sections. The present invention provides the above features in a system that is virtually transparent to the program code and does not require a variety of complex or specialized instructions or address coding methods. The present invention is flexible in that the two set associative instruction cache is transformed into what may be thought of as a static RAM in cache, and in addition, a direct map cache unit. Several different time critical code sections may be loaded and locked into the cache at different times.

Apparatus and method for an instruction cache locking scheme

A partially blocking data cache having improved microprocessor performance while maintaining data consistency between external memory and cache memory. The data cache of the present invention is used in a computer system and is partially blocking in that this cache will block the execution of any store instructions subsequent to an outstanding load instruction that missed the cache. The present invention offers increased microprocessor efficiency by allowing execution of subsequent load instructions while less than a predetermined number of preceding load instructions are still outstanding. The present invention utilizes a counter within the data cache unit to track the number of outstanding load misses. The present invention provides increased performance without undue or overly complex modifications to existing caching systems. The present invention operates advantageously within a computer system having a relatively large number of registers associated with the microprocessor such that store instructions represent a relatively small number of the instructions executed by the microprocessor.

Apparatus and method for partial execution blocking of instructions

A hierarchical memory which includes a backing store read/write memory (18) for storing first words, and a read-only memory RAM (60) for storing frequently used words. The buffer store has two parts, a cache RAM (64) and a two-word queue (62) comprised of two fetch buffers. The cache RAM is provided for storing a copy of some of the word stored in the backing store in accordance with a use algorithm. The ROM, queue buffers and cache RAM are simultaneously searched to see if the address for requested words is in either of them. If not, a fetch (76) is made of the backing store (18) and the words are written into the fetch buffers. The next time that address is presented, the fetch buffers are written into the cache and simultaneously read out to the bus. A first Y-mux (63) is provided between the ROM and the cache RAM for multiplexing the appropriate ROM columns to drive the Cache RAM bit lines directly when an internal micro-address is selected. The word positions within a row of the ROM are so ordered as to enable multiple words to be read out simultaneously regardless of what starting address is presented to the ROM. A second Y-mux (67) is provided between the cache RAM and the bus (19) for multiplexing the multiple words. The second Y-mux is controlled by at least one bit of the internally selected address. A multiplexer (74) is connected to the second Y-mux for shifting its input an appropriate amount according to bits of the starting address, such that the correct words are read out in the correct position.

High bandwith output hierarchical memory store including a cache, fetch

A functional-redundancy checking logic for checking identical chips whose outputs are time-variant programmable. Window logic (40, 60) on each chip creates a window associated with each programmed transition, set pulse (14) and reset pulse (15). For the set pulse (14), on the rising edge (42) of the window, an error flip-flop (50) is set. The flip-flop is merely set at this time; an error is not flagged. The window remains open for a fixed time period. During the time period that the window is open, if the output pin (24) is ever correctly asserted, then the flip-flop (50) is reset, thus clearing the error flag. However, if the pin (24) always remains incorrectly asserted, indicating an error, then the error flip-flop (50) remains set. When the window closes on the falling edge (46) of the window pulse, an error report pulse is created via the AND (52). The value on the flip-flop (50) at this time is reported as an error (54). An identical circuit checks the programmable reset pulse ( 15).

Growing community of inventors

Portland, OR, United States of America

Sunil R Shenoy