International Business Machines Corporation

Intel Corporation

Larry Edward Thatcher

David James Shippy

Hung Qui Le

Bruce Joseph Ronchetti

James Allan Kahle

Steven Wayne White

John Stephen Muhich

Rajesh Bhikhubhai Patel

Bryan D Boatright

Marlin Wayne Frederick, Jr

Kurt Alan Feiste

Brian R Konigsburg

John Edward Derrick

Troy N Hicks

T W Griffith

George McNeil Lattimore

Kevin Franklin Reick

Gus Wai-Yan Yeung

Robert Alan Cargnoni

Jose Angel Paredes

John Beck

David Scott Ray

Edward John Silha

Lance E Hacking

Sankaran M Menon

Charles Roberts Moore

Sanjoy K Mondal

Michael Kevin Ciraula

Robert Greg McDonald

Albert Y Chang

Srinivas Patil

David M Wu

Talal K Jaber

Henry Stanley Warren, Jr

Humberto Felipe Casal

Anil K Sabbavarapu

Richard E Fry

Chih-Jen M Lin

Eric L Hendrickson

Myhong Nguyenphu

William Hodges

Dave Shippy

Madhukar K Reddy

A technique to promote determinism among multiple clocking domains within a computer system or integrated circuit, In one embodiment, one or more execution units are placed in a deterministic state with respect to multiple clocks within a processor system having a number of different clocking domains.

Technique for promoting determinism among multiple clock domains

A testing architecture for testing a complex integrated circuit in which each functional unit may be tested independently of the others. Embodiments of the invention allow testing of functional units to take place at slower or faster clock speeds than other portions of the processor without incurring delay or other adverse timing effects.

Flexible scan architecture

A microprocessor, data processing system, and method are disclosed for handling parity errors in an address translation facility such as a TLB. The microprocessor includes a load/store unit configured to generate an effective address associated with a load/store instruction. An address translation unit adapted to translate the effective address to a real address using a translation lookaside buffer (TLB). The address translation unit includes a parity checker configured to verify the parity of the real address generated by the TLB and to signal the load store unit when the real address contains a parity error. The load store unit is configured to initiate a TLB parity error interrupt routine in response to the signal from the translation unit. In one embodiment, the TLB interrupt routine selectively invalidates the TLB entry that contained the parity error. The load/store unit preferably includes an effective to real address table (ERAT) containing a set of address translations. In this embodiment, the load/store unit invokes the address translation unit to translate the effective address only if the effective address misses in the ERAT. The LSU may suitably include an ERAT miss queue (EMQ) adapted to retain an effective address that misses in the ERAT until the address translation unit completes the translation process. In this embodiment, the EMQ is configured to issue a TLB parity error interrupt signal to initiate the TLB parity error interrupt routine. In one embodiment, the TLB interrupt routine loads a data address register (DAR) of the microprocessor with the effective address of the instruction that resulted in the parity error. The TLB interrupt routine may further set a data storage interrupt routine status register (DSISR) to indicate the TLB parity error.

TLB parity error recovery

The present invention relates to the use of multiple store buffer forwarding in a microprocessor system with a restrictive memory model. In accordance with an embodiment of the present invention, the system and method allow load operations that are completely covered by two or more store operations to receive data via store buffer forwarding in such a manner as to retain the side effects of the restrictive memory model thereby increasing processor performance without violating the restrictive memory model. In accordance with an embodiment the present invention, a method for multiple store buffer forwarding in a system with a restrictive memory model includes executing multiple store instructions, executing a load instruction, determining that a memory region addressed by the load instruction matches a cacheline address in a memory, determining that data stored by the multiple store instructions completely covers the memory region addressed by the load instruction, and transmitting a store forward is OK signal.

System and method for multiple store buffer forwarding in a system with a restrictive memory model

A microprocessor and related method and data processing system are disclosed. The microprocessor includes a dispatch unit suitable for issuing an instruction executable by the microprocessor, an execution pipeline configured to receive the issued instruction, and a pending instruction unit. The pending instruction unit includes a set of pending instruction entries. A copy of the issued instruction is maintained in one of the set of pending instruction entries. The execution pipeline is adapted to record, in response detecting to a condition preventing the instruction from successfully completing one of the stages in the pipeline during a current cycle, an exception status with the copy of the instruction in the pending instruction unit and to advance the instruction to a next stage in the pipeline in the next cycle thereby preventing the condition from stalling the pipeline. Preferably, the dispatch unit, in response to the instruction finishing pipeline execution with an exception status, is adapted to use the copy of the instruction to re-issue the instruction to the execution pipeline in a subsequent cycle. In one embodiment, the dispatch unit is adapted to deallocate the copy of the instruction in the pending instruction unit in response to the instruction successfully completing pipeline execution. The pending instruction unit may detect successful completion of the instruction by detecting when the instruction has been pending for a predetermined number of cycles without recording an exception status. In this embodiment, each entry in the pending instruction unit may include a timer field comprising a set of bits wherein the number of bits in the time field equals the predetermined number of cycles. The pending instruction unit may set, in successive cycles, successive bits in the timer field such that successful completion of an instruction is indicated when a last bit in the time field is set. In one embodiment, pending instruction unit includes a set of copies of instructions corresponding to each of a set of instructions pending in the execution pipeline at any given time. In various embodiments, the execution pipeline may comprise a load/store pipeline, a floating point pipeline, or a fixed point pipeline.

Determining successful completion of an instruction by comparing the number of pending instruction cycles with a number based on the number of stages in the pipeline

The present invention relates to locked memory instructions, and more specifically to a system and method for the high performance execution of locked memory instructions in a system with distributed memory and a restrictive memory model. In accordance with an embodiment of the present invention, a method for executing locked-memory instructions includes decoding a locked-memory instruction, obtaining exclusive ownership of a cacheline to be used by a load-lock operation, setting a bit to indicate the load-lock operation's ownership of the cacheline, and activating a snoop checking process. The method also includes modifying a load data value and storing the modified load data value. The method further includes determining that the cacheline is still exclusively owned, storing the load data value, determining that the cacheline is unsnooped, merging the modified load data value with the load data value, and releasing the locked-memory instruction to be retired.

System and method for high performance execution of locked memory instructions in a system with distributed memory and a restrictive memory model

A processor and an associated method and data processing system are disclosed. The processor includes an issue unit (ISU), a completion unit, and a hang detect unit. The ISU is configured to issue instructions to an execution unit. The completion unit is adapted to produce a completion valid signal responsive to the issue unit completing an instruction. The hang detect unit is configured to receive the completion valid signal from the ISU and adapted to determine the interval since the most recent assertion of the completion valid signal. The hang detect unit is adapted to initiate a hang recovery sequence upon determining that the interval since the most recent assertion of the completion valid signal exceeds a predetermined maximum interval. In one embodiment, the hang recovery sequence includes the hang recovery unit asserting a stop completion signal to a completion unit and a stop dispatch signal to a dispatch unit to suspend instruction completion and dispatch. The hang recovery unit then asserts a force reject signal to an execution unit to reject all instructions pending in the execution unit's pipeline and a flush signal to the execution unit that results in the processor flushing a set of instructions. The hang recovery unit then negates the force reject, stop completion, and stop dispatch signals to resume processor operation. In one embodiment, the recovery sequence includes entering a relaxed execution mode, such as a debug mode, a serial operation mode, or an in-order mode prior to resuming processor operation. In one embodiment, the processor advances a completion tag upon completing an instruction. In this manner the completion tag indicates the instruction that is next to complete. In one embodiment, the hang recovery sequence includes flushing the processor of an instruction set comprising all instructions with tag information greater than the completion tag. In another embodiment, all instructions with tag information greater than or equal to the completion tag are flushed.

Recovery from hang condition in a microprocessor

A method for optimally issuing instructions that are related to a first instruction in a data processing system is disclosed. The processing system includes a primary and secondary cache. The method and system comprises speculatively indicating a hit of the first instruction in a secondary cache and releasing the dependent instructions. The method and system includes determining if the first instruction is within the secondary cache. The method and system further includes providing data related to the first instruction from the secondary cache to the primary cache when the instruction is within the secondary cache. A method and system in accordance with the present invention causes instructions that create dependencies (such as a load instruction) to signal an issue queue (which is responsible for issuing instructions with resolved conflicts) in advance, that the instruction will complete in a predetermined number of cycles. In an embodiment, a core interface unit (CIU) will signal an execution unit such as the Load Store Unit (LSU) that it is assumed that the instruction will hit in the L cache. An issue queue uses the signal to issue dependent instructions at an optimal time. If the instruction misses in the L cache, the cache hierarchy causes the instructions to be abandoned and re-executed when the data is available.

Method and system for optimally issuing dependent instructions based on speculative L2 cache hit in a data processing system

A data processing system including a processor having a load/store unit and a method for correcting effective address aliasing. In the load/store unit within the processor, load and store instructions are executed out of order. The load and store instructions are assigned tags in a predetermined manner, and then assigned to load and store reorder queues for keeping track of the program order of the load and store instructions. A real address tag is utilized to correct for effective address aliasing within the load/store unit.

Data processing system including load/store unit having a real address tag array and method for correcting effective address aliasing

A method and system for optimizing execution of an instruction stream which includes a very long instruction word (VLIW) dispatch group in which ordering is not maintained is disclosed. The method and system comprises examining an access which initiated a flush operation; capturing an indice related to the flush operation; and causing all storage access instructions related to this indice to be dispatched as single-IOP groups until the indice is updated. Storage access to address space which is safe such as Guarded (G&equals;1) or Direct Store (E&equals;DS) must be handled in a non-speculative manner such that operations which could potentially go to volatile I/O devices or control locations that do not get processed out of order. Since the address is not known in the front end of the processor, this can only be determined by the load store unit or functional block which performs translation. Therefore, if a flush occurs for these conditions, in accordance with the present invention the value of the base register (RA) is latched and subsequent loads and stores which use this base register are decoded in a &ldquo;safe&rdquo; manner until an instruction is decoded which would change the base register value (safe means an internal instruction sequence which can be executed in order without repeating any accesses). The value of multiple base registers can be tracked in this manner, though the preferred embodiment would not use more than two, one of the base registers could be for input and one could be for output streams.

Optimization of instruction stream execution that includes a VLIW dispatch group

As a program is replaced by the operating system running within a microprocessor, only those entries associated with the replaced program and resident within effective-to-real address translation units will be replaced. Those entries within the effective-to-real address translation units associated with the operating system and shared libraries, and any other software units operating within the microprocessor will not be invalidated.

System and method for invalidating an entry in a translation unit

In a processor, store instructions are divided or cracked into store data and store address generation portions for separate and parallel execution within two execution units. The address generation portion of the store instruction is executed within the load store unit, while the store data portion of the instruction is executed in an execution unit other than the load store unit. If the store instruction is a fixed point execution unit, then the store data portion is executed within the fixed point unit. If the store instruction is a floating point store instruction, then the store data portion of the store instruction is executed within the floating point unit.

System and method for executing store instructions

Pipelining and parallel execution of multiple load instructions is performed within a load store unit. When a first load instruction incurs a cache miss and proceeds to retrieve the load data from the system memory hierarchy, a second load instruction addressing the same load data will be merged into the first load instruction so that the data returned from the system memory hierarchy is sent to register files associated with both the first and second load instructions. As a result, the second load instruction does not have to wait until the load data has been written and validated in the data cache.

System and method for merging multiple outstanding load miss instructions

In a load/store unit within a microprocessor, load and store instructions are executed out of order. The load and store instructions are assigned tags in a predetermined manner, and then assigned to load and store reorder queues for keeping track of the program order of the load and store instructions. Then when new load or store instructions are issued, the new load or store instructions are compared to entries within the load and store reorder queues to detect out of order problems.

System and method for permitting out-of-order execution of load and store instructions

Growing community of inventors

Austin, TX, United States of America

Larry Edward Thatcher