Patent US5778208A successfully refreshed from Google Patents.
US5778208A
Expired / Lapsed
US
⏱️ 0.0 years remaining
Flexible pipeline for interlock removal
Assignee: International Business Machines Corp
•
Filing: 1995-12-18
•
Granted: 1998-07-07
•
Expires: 2015-12-18
Abstract
Abstract A flexible pipeline for reducing performance limiting pipeline interlocks in the execution of programs. The pipeline architecture includes for each pipeline a fetch stage, a decode stage, an execution stage, a hybrid memory/execution stage, and a write back stage. When the result from the execution stage of a first pipeline is not available to a second pipeline until the write back stage of the first pipeline as a consequence of an interlock, the execution stage of the second pipeline may be delayed at least one execution cycle so that the executable functions are performed in the hybrid memory/execution stage or fourth stage of the second pipeline. The result from the execution stage is obtained either by a calculation of the effective address of a memory location or by performing arithmetic/logical unit (ALU) functions. The third stage of operation of the second pipeline carries out a null operation when the execute operation is performed in the fourth stage as a consequence of, for example, a load instruction being followed by a use instruction thereby resulting in a pipeline interlock.
Inventors
- 👤 Richard James Eickemeyer
- 👤 Nadeem Malik
- 👤 Avijit Saha
Bibliographic Details
Application Number
—
Jurisdiction
US
Data Source
google_patents
Last Synchronized
2026-11-08 01:41
CPC Classifications
G; G06; G06F; G06F9/00; G06F9/06; G06F9/30; G06F9/38; G06F9/3824
Claims Analysis
Total Claims: 45
Claim 1
Independent Claim
122 words
A pipeline architecture wherein execution of an operation during an execution cycle, the pipeline architecture comprising: a first execution unit for performing a first and second type of executable function in a first execution cycle; and a second execution unit for performing the first type of executable function and a third type of executable function in a second execution cycle, the second execution unit being offset in the pipeline by at least one execution cycle from the first execution unit, the offset of the second execution unit allowing immediate execution of the first type of executable function during the second execution cycle by the second execution unit when the first type of executable function is not performed in the first execution unit.
Claim 2
Dependent Claim
122 words
A pipeline architecture wherein execution of an operation during an execution cycle, the pipeline architecture comprising: a first execution unit for performing a first and second type of executable function in a first execution cycle; and a second execution unit for performing the first type of executable function and a third type of executable function in a second execution cycle, the second execution unit being offset in the pipeline by at least one execution cycle from the first execution unit, the offset of the second execution unit allowing immediate execution of the first type of executable function during the second execution cycle by the second execution unit when the first type of executable function is not performed in the first execution unit.
Claim 3
Dependent Claim
14 words
The pipeline architecture of claim 1 wherein data is retrieved from outside the pipeline.
Claim 4
Dependent Claim
15 words
The pipeline architecture of claim 1 wherein the second execution unit is paired with memory.
Claim 5
Dependent Claim
14 words
The pipeline architecture of claim 1 wherein the execution cycle comprises a clock cycle.
Claim 6
Dependent Claim
25 words
The pipeline architecture of claim 1 wherein the first execution unit carries out a null operation when execution is performed by the second execution unit.
Claim 7
Dependent Claim
26 words
The pipeline architecture of claim 1 further comprising means for making a selection between the use of the first execution unit and the second execution unit.
Claim 8
Dependent Claim
147 words
A pipeline architecture for executing serially a sequence of instructions in a succession of execution cycles, wherein a result is produced representing concurrent execution of a first instruction and a second instruction in which the second instruction requires a result produced by execution of the first instruction, the pipeline architecture comprising: a first pipeline comprising an execution stage, a hybrid memory/execution stage, and a write back stage; and a second pipeline, offset in operation from the first pipeline by one execution cycle, the second pipeline comprising an execution stage, a hybrid memory/execution stage, and a write back stage; wherein, as a consequence of an interlock, a result from the execution stage of the first pipeline is unavailable to the second pipeline until the write back stage of the first pipeline, and execution by the second pipeline is performed in the hybrid memory/execution stage of the second pipeline.
Claim 9
Dependent Claim
147 words
A pipeline architecture for executing serially a sequence of instructions in a succession of execution cycles, wherein a result is produced representing concurrent execution of a first instruction and a second instruction in which the second instruction requires a result produced by execution of the first instruction, the pipeline architecture comprising: a first pipeline comprising an execution stage, a hybrid memory/execution stage, and a write back stage; and a second pipeline, offset in operation from the first pipeline by one execution cycle, the second pipeline comprising an execution stage, a hybrid memory/execution stage, and a write back stage; wherein, as a consequence of an interlock, a result from the execution stage of the first pipeline is unavailable to the second pipeline until the write back stage of the first pipeline, and execution by the second pipeline is performed in the hybrid memory/execution stage of the second pipeline.
Claim 10
Dependent Claim
30 words
The pipeline architecture of claim 7 wherein the result is obtained either by a calculation of the effective address of a memory location or by performing arithmetic/logical unit (ALU) functions.
Claim 11
Dependent Claim
14 words
The pipeline architecture of claim 7 wherein the execution cycle comprises a clock cycle.
Claim 12
Dependent Claim
27 words
The pipeline architecture of claim 9 further comprising means for making a selection between the use of memory functions or execution functions in the hybrid memory/execution stage.
Claim 13
Dependent Claim
16 words
The pipeline architecture of claim 10 wherein the selection is accomplished prior to the execution stage.
Claim 14
Dependent Claim
54 words
The pipeline architecture of claim 7 wherein the execution stage of the second pipeline carries out a null operation in the execution stage when execution by the second pipeline is performed in the hybrid memory/execution stage as a consequence of a load instruction being followed by a use instruction resulting in a pipeline interlock.
Claim 15
Dependent Claim
34 words
The pipeline architecture of claim 7 wherein the first pipeline is a load pipeline and the second pipeline is an ALU pipeline, and the pipeline architecture further comprises a bypass of the ALU pipeline.
Claim 16
Dependent Claim
55 words
The pipeline architecture of claim 7 wherein the first and second pipelines function as ALU pipelines having two dependent ALU instructions dispatched at the same time, the first ALU instruction being executed in the execution stage of the first pipeline and the second ALU instruction being executed in the hybrid stage of the second pipeline.
Claim 17
Dependent Claim
167 words
A microprocessor system comprising: a bus for routing data in the microprocessor system; memory, coupled to the bus, for storing data; cache, coupled to the bus, for holding recently-accessed data; and a cpu module, coupled to the bus, the cpu module comprising at least one microprocessor; wherein the microprocessor performs pipeline operations according to a pipeline architecture, the pipeline architecture comprising: a first execution unit for performing a first and second type of executable function in a first execution cycle; and a second execution unit for performing the first type of executable function and a third type of executable function in a second execution cycle, the second execution unit being offset in the pipeline by at least one execution cycle from the first execution unit, the offset of the second execution unit allowing immediate execution of the first type of executable function during the second execution cycle by the second execution unit when the first type of executable function is not performed in the first execution unit.
Claim 18
Dependent Claim
167 words
A microprocessor system comprising: a bus for routing data in the microprocessor system; memory, coupled to the bus, for storing data; cache, coupled to the bus, for holding recently-accessed data; and a cpu module, coupled to the bus, the cpu module comprising at least one microprocessor; wherein the microprocessor performs pipeline operations according to a pipeline architecture, the pipeline architecture comprising: a first execution unit for performing a first and second type of executable function in a first execution cycle; and a second execution unit for performing the first type of executable function and a third type of executable function in a second execution cycle, the second execution unit being offset in the pipeline by at least one execution cycle from the first execution unit, the offset of the second execution unit allowing immediate execution of the first type of executable function during the second execution cycle by the second execution unit when the first type of executable function is not performed in the first execution unit.
Claim 19
Dependent Claim
15 words
The microprocessor system of claim 15 wherein the data is retrieved from outside the pipeline.
Claim 20
Dependent Claim
17 words
The microprocessor system of claim 15 wherein the second execution unit is paired with memory registers operation.
Claim 21
Dependent Claim
14 words
The microprocessor system of claim 15 wherein the execution cycle comprises a clock cycle.
Claim 22
Dependent Claim
25 words
The microprocessor system of claim 15 wherein the first execution unit carries out a null operation when execution is performed by the second execution unit.
Claim 23
Dependent Claim
26 words
The microprocessor system of claim 15 further comprising means for making a selection between the use of the first execution unit and the second execution unit.
Claim 24
Dependent Claim
162 words
A pipeline architecture, comprising: a first stage for fetching instructions from program text; a second stage, coupled to the first stage, for decoding the fetched instructions; a third stage, coupled to the second stage, the third stage having a first and a second operation mode, the first operation mode being an execution of the decoded instructions to produce a result in the absence of a load-use interlock instruction and the second operation mode being a null operation in the presence of a load-use interlock instruction; a fourth stage, coupled to the third stage, the fourth stage having a normal and a interlock operation mode, the normal operation mode being the fetching of data from a memory address in the absence of an interlock instruction, and the interlock operation being the execution of the decoded instructions to produce a result in the presence of the interlock instruction; and a fifth stage, coupled to the fourth stage, for writing the result to a register.
Claim 25
Dependent Claim
162 words
A pipeline architecture, comprising: a first stage for fetching instructions from program text; a second stage, coupled to the first stage, for decoding the fetched instructions; a third stage, coupled to the second stage, the third stage having a first and a second operation mode, the first operation mode being an execution of the decoded instructions to produce a result in the absence of a load-use interlock instruction and the second operation mode being a null operation in the presence of a load-use interlock instruction; a fourth stage, coupled to the third stage, the fourth stage having a normal and a interlock operation mode, the normal operation mode being the fetching of data from a memory address in the absence of an interlock instruction, and the interlock operation being the execution of the decoded instructions to produce a result in the presence of the interlock instruction; and a fifth stage, coupled to the fourth stage, for writing the result to a register.
Claim 26
Dependent Claim
30 words
The pipeline architecture of claim 21 wherein the result is obtained either by a calculation of the effective address of a memory location or by performing arithmetic/logical unit (ALU) functions.
Claim 27
Dependent Claim
46 words
The pipeline architecture of claim 21 wherein the third stage carries out a null operation when the execution of the decoded instructions is performed in the fourth stage as a consequence of a load instruction being followed by a use instruction resulting in a pipeline interlock.
Claim 28
Dependent Claim
25 words
The pipeline architecture of claim 21 further comprising means for making a selection between the normal and the interlock operation modes of the fourth stage.
Claim 29
Dependent Claim
15 words
The pipeline architecture of claim 24 wherein the selection is accomplished at the second stage.
Claim 30
Dependent Claim
150 words
An instruction pipeline having a plurality of serially operating instruction stages for processing a plurality of overlapping instructions during each execution cycle, comprising: fetch means, operable in a first stage, for fetching instructions from program text; decode means, operable in a second stage coupled to the first stage, for receiving the fetched instruction and determining the functions to be performed on operands in the instruction, the instruction and operands being converted into an internal instruction format; execute means, operable in a third or in a fourth stage, for performing executable functions based upon the decoded instructions and operands; memory means, operable in the fourth stage, for fetching data from a memory address; and write back means for writing a result obtained from the execution functions to a register; wherein the performance of the executable functions is delayed one cycle to the fourth stage in the presence of an interlock instruction.
Claim 31
Dependent Claim
150 words
An instruction pipeline having a plurality of serially operating instruction stages for processing a plurality of overlapping instructions during each execution cycle, comprising: fetch means, operable in a first stage, for fetching instructions from program text; decode means, operable in a second stage coupled to the first stage, for receiving the fetched instruction and determining the functions to be performed on operands in the instruction, the instruction and operands being converted into an internal instruction format; execute means, operable in a third or in a fourth stage, for performing executable functions based upon the decoded instructions and operands; memory means, operable in the fourth stage, for fetching data from a memory address; and write back means for writing a result obtained from the execution functions to a register; wherein the performance of the executable functions is delayed one cycle to the fourth stage in the presence of an interlock instruction.
Claim 32
Dependent Claim
28 words
The instruction pipeline of claim 26 wherein the executable functions comprise a calculation of the effective address of a memory location or execution of arithmetic/logical unit (ALU) functions.
Claim 33
Dependent Claim
14 words
The instruction pipeline of claim 26 wherein the execution cycle comprises a clock cycle.
Claim 34
Dependent Claim
42 words
The instruction pipeline of claim 26 wherein the third stage performs a null operation when the executable functions are performed in the fourth stage as a consequence of a load instruction being followed by a use instruction resulting in a pipeline interlock.
Claim 35
Dependent Claim
28 words
The instruction pipeline of claim 26 further comprising means for making a selection between fetching data from a memory address or performing executable functions in the fourth stage.
Claim 36
Dependent Claim
15 words
The instruction pipeline of claim 26 wherein the selection is accomplished at the second stage.
Claim 37
Independent Claim
125 words
A method of processing serially operating instruction during an execution cycle, wherein execution of an operation during an execution cycle depends upon retrieval of data from outside the pipeline, the method comprising the steps of: performing a first and second type of executable function in a first execution unit on data being advanced through the pipeline; and performing the first type of executable function and a third type of executable function in a second execution unit offset by at least one execution cycle from the first execution unit, the first type of executable function being performed immediately in the second execution unit upon retrieval of data from outside the pipeline when the first type of executable function is not performed in the first execution unit.
Claim 38
Independent Claim
125 words
A method of processing serially operating instruction during an execution cycle, wherein execution of an operation during an execution cycle depends upon retrieval of data from outside the pipeline, the method comprising the steps of: performing a first and second type of executable function in a first execution unit on data being advanced through the pipeline; and performing the first type of executable function and a third type of executable function in a second execution unit offset by at least one execution cycle from the first execution unit, the first type of executable function being performed immediately in the second execution unit upon retrieval of data from outside the pipeline when the first type of executable function is not performed in the first execution unit.
Claim 39
Dependent Claim
30 words
The method of claim 32 further comprising the step of performing a null operation in the first execution unit when execution on data is performed by the second execution unit.
Claim 40
Dependent Claim
24 words
The method of claim 32 further comprising the step of selecting between the use of the first execution unit and the second execution unit.
Claim 41
Independent Claim
165 words
A method of processing serially operating instruction per execution cycle, comprising the steps of: fetching instructions from program text during a first stage; obtaining, in a second stage, operation codes from the fetched instructions and determining the functions to be performed on operands, the codes and functions being converted into an internal instruction format; determining whether a load-use interlock will occur; performing executable functions in a third stage based upon the decoded instructions and operands as a result of determining that a load-use interlock will not occur and inserting a null in the third stage as a result of determining that a load-use interlock will occur; performing, in a fourth stage, executable functions based upon the decoded instructions and operands as a result of determining that an interlock will occur and fetching data from a memory address in the fourth stage as a result of determining that an interlock use will not occur; and writing a result obtained from the executable functions to a register.
Claim 42
Independent Claim
165 words
A method of processing serially operating instruction per execution cycle, comprising the steps of: fetching instructions from program text during a first stage; obtaining, in a second stage, operation codes from the fetched instructions and determining the functions to be performed on operands, the codes and functions being converted into an internal instruction format; determining whether a load-use interlock will occur; performing executable functions in a third stage based upon the decoded instructions and operands as a result of determining that a load-use interlock will not occur and inserting a null in the third stage as a result of determining that a load-use interlock will occur; performing, in a fourth stage, executable functions based upon the decoded instructions and operands as a result of determining that an interlock will occur and fetching data from a memory address in the fourth stage as a result of determining that an interlock use will not occur; and writing a result obtained from the executable functions to a register.
Claim 43
Dependent Claim
25 words
The method of claim 35 wherein the step of performing executable functions further comprises the step of calculating the effective address of a memory location.
Claim 44
Dependent Claim
22 words
The method of claim 35 wherein the step of performing executable functions further comprises the step of executing arithmetic/logical unit (ALU) functions.
Claim 45
Dependent Claim
34 words
The method of claim 35 further comprising the step of making the third stage a null operation when the execute operation is performed in the fourth stage as a result of an interlock occurring.
References Cited (Backward Citations)
US4750112A
Cited
USRE34052E
Cited
US5081575A
Cited
US5099421A
Cited
US4985825A
Cited
US5297281A
Cited
US5051940A
Cited
US5299319A
Cited
US5193167A
Cited
JPH04105126A
Cited
US5333284A
Cited
US5367648A
Cited
US5404552A
Cited
US5426743A
Cited
US5434993A
Cited
US5416911A
Cited
USPTO Annuity & Maintenance Windows
| Annuity Stage | Window Opens | Payment Deadline | Surcharge Period Until | Estimated Fee | Status |
|---|---|---|---|---|---|
| 3.5 Years (Year 4) | 2001-07-06 | 2002-01-05 | 2002-07-07 | $2000.00 | Window Closed / Lapsed |
| 7.5 Years (Year 8) | 2005-07-06 | 2006-01-05 | 2006-07-07 | $3760.00 | Window Closed / Lapsed |
| 11.5 Years (Year 12) | 2009-07-06 | 2010-01-05 | 2010-07-07 | $7700.00 | Window Closed / Lapsed |