WO2006086713A2 - Conditional instruction execution via emissary instruction for condition evaluation - Google Patents

Conditional instruction execution via emissary instruction for condition evaluation Download PDF

Info

Publication number
WO2006086713A2
WO2006086713A2 PCT/US2006/004892 US2006004892W WO2006086713A2 WO 2006086713 A2 WO2006086713 A2 WO 2006086713A2 US 2006004892 W US2006004892 W US 2006004892W WO 2006086713 A2 WO2006086713 A2 WO 2006086713A2
Authority
WO
WIPO (PCT)
Prior art keywords
instruction
condition
emissary
conditional
base
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Ceased
Application number
PCT/US2006/004892
Other languages
French (fr)
Other versions
WO2006086713A3 (en
Inventor
Michael Scott Mcilvaine
James Norris Dieffenderfer
Jeffrey Todd Bridges
Thomas Andrew Sartorius
Rodney Wayne Smith
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Qualcomm Inc
Original Assignee
Qualcomm Inc
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Qualcomm Inc filed Critical Qualcomm Inc
Priority to JP2007555283A priority Critical patent/JP4903720B2/en
Priority to DE602006013415T priority patent/DE602006013415D1/en
Priority to EP06734847A priority patent/EP1846817B1/en
Priority to BRPI0606973-8A priority patent/BRPI0606973A2/en
Priority to AT06734847T priority patent/ATE463787T1/en
Publication of WO2006086713A2 publication Critical patent/WO2006086713A2/en
Publication of WO2006086713A3 publication Critical patent/WO2006086713A3/en
Priority to IL185066A priority patent/IL185066A0/en
Anticipated expiration legal-status Critical
Ceased legal-status Critical Current

Links

Classifications

    • GPHYSICS
    • G06COMPUTING OR CALCULATING; COUNTING
    • G06FELECTRIC DIGITAL DATA PROCESSING
    • G06F9/00Arrangements for program control, e.g. control units
    • G06F9/06Arrangements for program control, e.g. control units using stored programs, i.e. using an internal store of processing equipment to receive or retain programs
    • G06F9/30Arrangements for executing machine instructions, e.g. instruction decode
    • G06F9/30003Arrangements for executing specific machine instructions
    • G06F9/30072Arrangements for executing specific machine instructions to perform conditional operations, e.g. using predicates or guards
    • GPHYSICS
    • G06COMPUTING OR CALCULATING; COUNTING
    • G06FELECTRIC DIGITAL DATA PROCESSING
    • G06F9/00Arrangements for program control, e.g. control units
    • G06F9/06Arrangements for program control, e.g. control units using stored programs, i.e. using an internal store of processing equipment to receive or retain programs
    • G06F9/30Arrangements for executing machine instructions, e.g. instruction decode
    • G06F9/3017Runtime instruction translation, e.g. macros
    • GPHYSICS
    • G06COMPUTING OR CALCULATING; COUNTING
    • G06FELECTRIC DIGITAL DATA PROCESSING
    • G06F9/00Arrangements for program control, e.g. control units
    • G06F9/06Arrangements for program control, e.g. control units using stored programs, i.e. using an internal store of processing equipment to receive or retain programs
    • G06F9/30Arrangements for executing machine instructions, e.g. instruction decode
    • G06F9/30181Instruction operation extension or modification
    • GPHYSICS
    • G06COMPUTING OR CALCULATING; COUNTING
    • G06FELECTRIC DIGITAL DATA PROCESSING
    • G06F9/00Arrangements for program control, e.g. control units
    • G06F9/06Arrangements for program control, e.g. control units using stored programs, i.e. using an internal store of processing equipment to receive or retain programs
    • G06F9/30Arrangements for executing machine instructions, e.g. instruction decode
    • G06F9/38Concurrent instruction execution, e.g. pipeline or look ahead
    • G06F9/3824Operand accessing
    • GPHYSICS
    • G06COMPUTING OR CALCULATING; COUNTING
    • G06FELECTRIC DIGITAL DATA PROCESSING
    • G06F9/00Arrangements for program control, e.g. control units
    • G06F9/06Arrangements for program control, e.g. control units using stored programs, i.e. using an internal store of processing equipment to receive or retain programs
    • G06F9/30Arrangements for executing machine instructions, e.g. instruction decode
    • G06F9/38Concurrent instruction execution, e.g. pipeline or look ahead
    • G06F9/3824Operand accessing
    • G06F9/3826Bypassing or forwarding of data results, e.g. locally between pipeline stages or within a pipeline stage
    • GPHYSICS
    • G06COMPUTING OR CALCULATING; COUNTING
    • G06FELECTRIC DIGITAL DATA PROCESSING
    • G06F9/00Arrangements for program control, e.g. control units
    • G06F9/06Arrangements for program control, e.g. control units using stored programs, i.e. using an internal store of processing equipment to receive or retain programs
    • G06F9/30Arrangements for executing machine instructions, e.g. instruction decode
    • G06F9/38Concurrent instruction execution, e.g. pipeline or look ahead
    • G06F9/3867Concurrent instruction execution, e.g. pipeline or look ahead using instruction pipelines

Definitions

  • the present invention relates generally to the field of processors and in particular to a method of conditional instruction execution via an emissary instruction for condition evaluation.
  • Microprocessors perform computational tasks in a wide variety of applications. High execution speed, low power consumption and small size are commonly important goals for processor designers, particularly in embedded applications such as portable electronic devices. An important consideration, however, is accuracy. Whatever architectural features are included in the processor design to improve execution speed or other metric needs also to assure the correct execution of software code.
  • conditional instructions are instructions that perform some operation - or not - depending on the evaluation of a condition.
  • Condition evaluation is typically performed in an execute pipe stage, deep in the pipeline.
  • Conditional instructions can greatly complicate hazard detection by delaying knowledge of whether the operation will be executed until deep in the pipeline.
  • knowledge of whether or not the operation will affect, e.g., the contents of a register is required early in the pipeline for hazard detection.
  • the following example code segment illustrates this problem: ADD rl, r7, r8 add contents of r7 and r8, place sum in rl
  • CMP rl, rl2 compare contents ofrl and rl2 (set code or flag to reflect the result of the comparison)
  • the ADDNE must calculate its result prior to the AND executing, so that the data written into rl by the ADDNE will be available to the AND.
  • the ADDNE is a conditional instruction, and may not execute the ADD operation.
  • the true dependency of the AND instruction is on the ADDEQ instruction.
  • exactly one of the ADDEQ and ADDNE instructions will execute, since the conditions EQ (equal) and NE (not equal) are mutually exclusive and exhaustive.
  • the AND since the ADDEQ instruction is conditional, the AND may actually have a dependency on an earlier instruction, in this case the initial ADD.
  • conditional instructions and in particular the possible multi-level contingent dependencies, greatly complicate the task of accurate hazard detection, which is necessary for accurately conducting out-of-order program execution.
  • a conditional instruction operative to perform an operation if a condition is satisfied, is converted to an emissary instruction operative to evaluate the condition and an unconditional base instruction operative to perform the operation.
  • the emissary instruction is executed, while the base instruction is halted.
  • the emissary instruction evaluates the condition and reports the condition evaluation back to the base instruction. Based on the condition evaluation, the base instruction is either launched into the pipeline for execution, or it is discarded (or a NOP, or null instruction, substituted for it). In either case, the dependencies of following instructions may be resolved.
  • a processor comprises an instruction execution pipeline including an instruction issue pipe stage and a condition evaluation pipe stage.
  • the processor includes a feedback path from the condition evaluation pipe stage to the instruction issue pipe stage.
  • the processor further includes a pipeline controller operative to convert a conditional instruction into an emissary instruction and an unconditional base instruction at the instruction issue pipe stage, issue the emissary instruction, and conditionally issue the base instruction in response to feedback from the condition evaluation pipe stage by the emissary instruction.
  • a conditional instruction operative to perform an operation only if a condition is satisfied, is decoded.
  • the conditional instruction is converted to an emissary instruction operative to evaluate the condition, and an unconditional base instruction operative to perform the operation.
  • the emissary instruction is executed. If the condition is not satisfied, it is determined that no other instruction is dependent on the conditional instruction.
  • a conditional instruction is decoded, where the conditional instruction is operative to perform an operation only if a condition is TRUE.
  • the conditional instruction is converted into two separate instructions, an emissary instruction, and an unconditional base instruction operative upon execution to perform the operation.
  • the emissary instruction is executed and caused to evaluate the condition.
  • the base instruction is executed or not depending upon the evaluation of the condition by the emissary instruction.
  • a data dependency between one or more following instructions and the base instruction is detected.
  • Another embodiment relates to a method of executing a conditional instruction.
  • the conditional instruction is converted into an emissary instruction and a base instruction.
  • the emissary instruction is executed.
  • the base instruction is conditionally executed in response to a condition evaluation by the emissary instruction.
  • Figure 1 is a functional block diagram of a processor.
  • Figure 2 is a functional block diagram of a superscalar pipeline.
  • Figure 3 is a flow diagram of conditional instruction execution.
  • FIG. 1 depicts a functional block diagram of a processor 10.
  • the processor 10 executes instructions in an instruction execution pipeline 12 according to control logic 14.
  • the pipeline 12 may be a superscalar design, with multiple parallel pipelines such as 12a and 12b.
  • the pipelines 12a, 12b include various registers or latches 16, organized in pipe stages, and one or more Arithmetic Logic Units (ALU) 18.
  • a General Purpose Register (GPR) file 20 provides registers comprising the top of the memory hierarchy.
  • the pipelines 12a, 12b fetch instructions from an Instruction Cache (I-Cache or 1$) 22, with memory addressing and permissions managed by an Instruction-side Translation Lookaside Buffer (ITLB) 24.
  • I-Cache Instruction Cache
  • ILB Instruction-side Translation Lookaside Buffer
  • Data is accessed from a Data Cache (D-Cache or D$) 26, with memory addressing and permissions managed by a main Translation Lookaside Buffer (TLB) 28.
  • the ITLB may comprise a copy of part of the TLB.
  • the ITLB and TLB may be integrated.
  • the I-cache 22 and D-cache 26 may be integrated, or unified. Misses in the I-cache 22 and/or the D-cache 26 cause an access to main (off-chip) memory 32, under the control of a memory interface 30.
  • the processor 10 may include an Input/Output (I/O) interface 34, controlling access to various peripheral devices 36.
  • I/O Input/Output
  • FIG. 1 depicts a superscalar pipeline architecture.
  • Processor pipeline 12 includes an instruction fetch pipe stage 40, fetching instructions from the I-Cache 22. The instructions are decoded at a decode pipe stage 42. Once the instruction is known, hazard detection logic 44 detects dependencies between the decoded instruction and instructions that are currently executing, such as instructions that generate operands needed by the decoded instruction. Following the hazard detection logic 44 is an expansion/issue pipe stage 46, that issues the decode instruction into one of multiple execution pipelines 12a, 12b.
  • conditional instructions may be transformed into two related instructions: an emissary instruction and a base instruction.
  • an "emissary instruction” is an instruction whose sole operation is the evaluation of the condition of a conditional instruction, and reporting that condition evaluation back to the base instruction.
  • the emissary instruction does not read or write any general purpose register, does not alter any condition code, and does not perform any memory access.
  • the emissary instruction corresponding to the conditional instruction "ADDEQ rl, r3, r4" is an instruction that detects and reports the EQ (equal) condition, e.g., by inspecting an EQ bit in a condition register that was set by a prior compare or other instruction.
  • the emissary instruction neither reads r3 or r4, nor writes rl.
  • a "base instruction” is an unconditional instruction that performs the same operation, with the same operands and destination, as the conditional instruction would perform if its condition evaluated TRUE.
  • the base instruction corresponding to the conditional instruction "ADDEQ rl, r3, r4" is "ADD rl, r3, r4".
  • the base instruction is halted at the issue pipe stage 46, pending the evaluation of its condition.
  • the emissary instruction is launched into an execution pipeline 12a, 12b, while the base instruction is halted.
  • the emissary instruction proceeds until it reaches a condition evaluation pipe stage 48, 50.
  • the emissary instruction evaluates the condition relevant to the conditional instruction, and reports the condition evaluation back to the base instruction halted at the issue pipe stage 46.
  • the emissary instruction may then proceed through the remaining pipeline 12a, 12b; may be replaced by a NOP (no-op or null instruction) that proceeds through the remaining pipeline 12a, 12b; or it may be discarded and the resulting pipe "bubble" filled by the next following instruction in the pipeline 12a, 12b.
  • NOP no-op or null instruction
  • the emissary instruction's useful life is over upon the evaluation and reporting of the conditional instruction's condition evaluation, and the emissary instruction may then be disposed of in any manner that does not affect the processor's program state, as well understood by those of skill in the art.
  • the base instruction halted at the issue pipe stage 46 is launched into an available execution pipeline 12a, 12b if the condition evaluation is TRUE.
  • a condition evaluation is considered TRUE if the condition is satisfied or a conditional test PASSes.
  • a condition evaluation is FALSE if the condition is not satisfied or a conditional test FAILs.
  • the hazard detection logic 44 may perform hazard detection for all instructions following the base instruction that read any register or memory location that the base instruction writes.
  • the base instruction is not executed.
  • the base instruction may be replaced by a NOP, which is launched into an available execution pipeline 12a, 12b.
  • the base instruction may be discarded, and a following instruction advanced to fill the pipeline bubble.
  • the hazard detection logic 44 knows that the base instruction (and correspondingly, the original conditional instruction) did not alter any register or memory location, and that no following instruction has any dependency on the original conditional instruction.
  • the hazard detection logic 44 may then look past the conditional instruction to one or more prior instructions, to determine a following instruction's actual dependency(s).
  • the base instruction is either discarded or converted to a NOP, which is issued into the pipeline 12a, 12b (block 64). In either case, and also if the decoded instruction is not a conditional instruction (block 54), hazard detection is performed for following instructions (block 66), as it is known whether or not the conditional instruction will execute.

Landscapes

  • Engineering & Computer Science (AREA)
  • Software Systems (AREA)
  • Theoretical Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • General Engineering & Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • Advance Control (AREA)
  • Debugging And Monitoring (AREA)
  • Control Of Transmission Device (AREA)
  • Semiconductor Integrated Circuits (AREA)
  • Electronic Switches (AREA)
  • Test And Diagnosis Of Digital Computers (AREA)

Abstract

Hazard detection is simplified by converting a conditional instruction, operative to perform an operation if a condition is satisfied, into an emissary instruction operative to evaluate the condition and an unconditional base instruction operative to perform the operation. The emissary instruction is executed, while the base instruction is halted. The emissary instruction evaluates the condition and reports the condition evaluation back to the base instruction. Based on the condition evaluation, the base instruction is either launched into the pipeline for execution, or it is discarded (or a NOP, or null instruction, substituted for it). In either case, the dependencies of following instructions may be resolved.

Description

CONDITIONAL INSTRUCTION EXECUTION VIA EMISSARY INSTRUCTION FOR CONDITION EVALUATION
BACKGROUND
[0001] The present invention relates generally to the field of processors and in particular to a method of conditional instruction execution via an emissary instruction for condition evaluation.
[0002] Microprocessors perform computational tasks in a wide variety of applications. High execution speed, low power consumption and small size are commonly important goals for processor designers, particularly in embedded applications such as portable electronic devices. An important consideration, however, is accuracy. Whatever architectural features are included in the processor design to improve execution speed or other metric needs also to assure the correct execution of software code.
[0003] Many modern processors employ a pipelined architecture, where sequential instructions, each having multiple execution steps, are overlapped in execution. For improved performance, the instructions should flow continuously through the pipeline. However, instructions may become stalled in the pipeline for a variety of reasons, such as data dependencies between instructions, also known as data hazards. High performance processors may employ a superscalar design, where part of the pipeline is replicated. Multiple pipelines can minimize the effect of pipeline stalls on overall system performance, as instructions awaiting a stalled pipeline may be dispatched to another pipeline for immediate execution.
[0004] To take maximum advantage of the potential performance gain offered by superscalar pipelines, high performance processors commonly support out-of-order execution. That is, instructions may be executed in other than their program order, if doing so is not precluded by inter-dependencies among the instructions. As one example of such dependencies, an instruction performing a logical or arithmetic operation that reads an operand from a register cannot execute until a prior instruction, that writes a result to that register, has calculated its result (data can be forwarded, where completed may imply the instruction has left the pipeline). These two instructions cannot be executed out of order, or an incorrect result will occur. The process of detecting dependencies among instructions is known as hazard detection. Hazard detection is commonly performed very early in the pipeline, prior to dispatching instructions into an execution pipeline.
[0005] Most instruction set architectures include conditional instructions, which are instructions that perform some operation - or not - depending on the evaluation of a condition. Condition evaluation is typically performed in an execute pipe stage, deep in the pipeline. Conditional instructions can greatly complicate hazard detection by delaying knowledge of whether the operation will be executed until deep in the pipeline. However, knowledge of whether or not the operation will affect, e.g., the contents of a register, is required early in the pipeline for hazard detection. The following example code segment illustrates this problem: ADD rl, r7, r8 add contents of r7 and r8, place sum in rl
CMP rl, rl2 compare contents ofrl and rl2 (set code or flag to reflect the result of the comparison)
ADDEQ rl, r8, r9 if the previous compare was equal, add contents ofr8 and r9 and place sum in rl. Otherwise, rl is not changed
ADDNE rl, r3, r4 if the previous compare was not equal, add contents ofrS and r4 and place sum in rl. Otherwise, rl is not changed
AND rlO, r6, rl logical AND the contents ofrό to the contents ofrl, place the result in rlO [0006] The final AND instruction has an rl dependency on the ADDNE instruction.
That is, the ADDNE must calculate its result prior to the AND executing, so that the data written into rl by the ADDNE will be available to the AND. However, the ADDNE is a conditional instruction, and may not execute the ADD operation. In this case, the true dependency of the AND instruction is on the ADDEQ instruction. In the foregoing example (depicting a common software construct), exactly one of the ADDEQ and ADDNE instructions will execute, since the conditions EQ (equal) and NE (not equal) are mutually exclusive and exhaustive. However, in general, since the ADDEQ instruction is conditional, the AND may actually have a dependency on an earlier instruction, in this case the initial ADD. Additionally, if a prior conditional instruction does not execute, a following instruction may have no dependency at all. [0007] The contingent dependencies presented by conditional instructions, and in particular the possible multi-level contingent dependencies, greatly complicate the task of accurate hazard detection, which is necessary for accurately conducting out-of-order program execution.
SUMMARY
[0008] In one embodiment, a conditional instruction, operative to perform an operation if a condition is satisfied, is converted to an emissary instruction operative to evaluate the condition and an unconditional base instruction operative to perform the operation. The emissary instruction is executed, while the base instruction is halted. The emissary instruction evaluates the condition and reports the condition evaluation back to the base instruction. Based on the condition evaluation, the base instruction is either launched into the pipeline for execution, or it is discarded (or a NOP, or null instruction, substituted for it). In either case, the dependencies of following instructions may be resolved. [0009] In another embodiment, a processor comprises an instruction execution pipeline including an instruction issue pipe stage and a condition evaluation pipe stage. The processor includes a feedback path from the condition evaluation pipe stage to the instruction issue pipe stage. The processor further includes a pipeline controller operative to convert a conditional instruction into an emissary instruction and an unconditional base instruction at the instruction issue pipe stage, issue the emissary instruction, and conditionally issue the base instruction in response to feedback from the condition evaluation pipe stage by the emissary instruction.
[0010] In another embodiment, a conditional instruction, operative to perform an operation only if a condition is satisfied, is decoded. The conditional instruction is converted to an emissary instruction operative to evaluate the condition, and an unconditional base instruction operative to perform the operation. The emissary instruction is executed. If the condition is not satisfied, it is determined that no other instruction is dependent on the conditional instruction.
[0011] Another embodiment relates to a method of performing hazard detection in a pipelined processor. A conditional instruction is decoded, where the conditional instruction is operative to perform an operation only if a condition is TRUE. The conditional instruction is converted into two separate instructions, an emissary instruction, and an unconditional base instruction operative upon execution to perform the operation. The emissary instruction is executed and caused to evaluate the condition. The base instruction is executed or not depending upon the evaluation of the condition by the emissary instruction. A data dependency between one or more following instructions and the base instruction is detected.
[0012] Another embodiment relates to a method of executing a conditional instruction. The conditional instruction is converted into an emissary instruction and a base instruction. The emissary instruction is executed. The base instruction is conditionally executed in response to a condition evaluation by the emissary instruction.
BRIEF DESCRIPTION OF DRAWINGS
[0013] Figure 1 is a functional block diagram of a processor.
[0014] Figure 2 is a functional block diagram of a superscalar pipeline.
[0015] Figure 3 is a flow diagram of conditional instruction execution.
DETAILED DESCRIPTION
[0016] Figure 1 depicts a functional block diagram of a processor 10. The processor 10 executes instructions in an instruction execution pipeline 12 according to control logic 14. The pipeline 12 may be a superscalar design, with multiple parallel pipelines such as 12a and 12b. The pipelines 12a, 12b include various registers or latches 16, organized in pipe stages, and one or more Arithmetic Logic Units (ALU) 18. A General Purpose Register (GPR) file 20 provides registers comprising the top of the memory hierarchy. The pipelines 12a, 12b fetch instructions from an Instruction Cache (I-Cache or 1$) 22, with memory addressing and permissions managed by an Instruction-side Translation Lookaside Buffer (ITLB) 24.
[0017] Data is accessed from a Data Cache (D-Cache or D$) 26, with memory addressing and permissions managed by a main Translation Lookaside Buffer (TLB) 28. In various embodiments, the ITLB may comprise a copy of part of the TLB. Alternatively, the ITLB and TLB may be integrated. Similarly, in various embodiments of the processor 10, the I-cache 22 and D-cache 26 may be integrated, or unified. Misses in the I-cache 22 and/or the D-cache 26 cause an access to main (off-chip) memory 32, under the control of a memory interface 30. The processor 10 may include an Input/Output (I/O) interface 34, controlling access to various peripheral devices 36. Those of skill in the art will recognize that numerous variations of the processor 10 are possible. For example, the processor 10 may include a second-level (L2) cache for either or both the I and D caches. In addition, one or more of the functional blocks depicted in the processor 10 may be omitted from a particular embodiment. [0018] Figure 2 depicts a superscalar pipeline architecture. Processor pipeline 12 includes an instruction fetch pipe stage 40, fetching instructions from the I-Cache 22. The instructions are decoded at a decode pipe stage 42. Once the instruction is known, hazard detection logic 44 detects dependencies between the decoded instruction and instructions that are currently executing, such as instructions that generate operands needed by the decoded instruction. Following the hazard detection logic 44 is an expansion/issue pipe stage 46, that issues the decode instruction into one of multiple execution pipelines 12a, 12b.
[0019] As discussed above, hazard detection is greatly complicated by conditional instructions. To simplify this task, the conditional instructions may be transformed into two related instructions: an emissary instruction and a base instruction. As used herein, an "emissary instruction" is an instruction whose sole operation is the evaluation of the condition of a conditional instruction, and reporting that condition evaluation back to the base instruction. The emissary instruction does not read or write any general purpose register, does not alter any condition code, and does not perform any memory access. For example, the emissary instruction corresponding to the conditional instruction "ADDEQ rl, r3, r4" is an instruction that detects and reports the EQ (equal) condition, e.g., by inspecting an EQ bit in a condition register that was set by a prior compare or other instruction. The emissary instruction neither reads r3 or r4, nor writes rl. [0020] As used herein, a "base instruction" is an unconditional instruction that performs the same operation, with the same operands and destination, as the conditional instruction would perform if its condition evaluated TRUE. For example, the base instruction corresponding to the conditional instruction "ADDEQ rl, r3, r4" is "ADD rl, r3, r4". In one embodiment, the base instruction is halted at the issue pipe stage 46, pending the evaluation of its condition.
[0021] Once the conditional instruction is transformed into an emissary instruction and a base instruction, the emissary instruction is launched into an execution pipeline 12a, 12b, while the base instruction is halted. The emissary instruction proceeds until it reaches a condition evaluation pipe stage 48, 50. There, the emissary instruction evaluates the condition relevant to the conditional instruction, and reports the condition evaluation back to the base instruction halted at the issue pipe stage 46. The emissary instruction may then proceed through the remaining pipeline 12a, 12b; may be replaced by a NOP (no-op or null instruction) that proceeds through the remaining pipeline 12a, 12b; or it may be discarded and the resulting pipe "bubble" filled by the next following instruction in the pipeline 12a, 12b. In short, the emissary instruction's useful life is over upon the evaluation and reporting of the conditional instruction's condition evaluation, and the emissary instruction may then be disposed of in any manner that does not affect the processor's program state, as well understood by those of skill in the art.
[0022] Upon reporting of the condition evaluation by the emissary instruction, the base instruction halted at the issue pipe stage 46 is launched into an available execution pipeline 12a, 12b if the condition evaluation is TRUE. A condition evaluation is considered TRUE if the condition is satisfied or a conditional test PASSes. Conversely, a condition evaluation is FALSE if the condition is not satisfied or a conditional test FAILs. In the case of a TRUE condition evaluation, the unconditional base instruction executes, and its effect on the register file or memory is known. Thus, the hazard detection logic 44 may perform hazard detection for all instructions following the base instruction that read any register or memory location that the base instruction writes. [0023] If the condition evaluation reported by the emissary instruction is FALSE, the base instruction is not executed. The base instruction may be replaced by a NOP, which is launched into an available execution pipeline 12a, 12b. Alternatively, the base instruction may be discarded, and a following instruction advanced to fill the pipeline bubble. In either case, the hazard detection logic 44 knows that the base instruction (and correspondingly, the original conditional instruction) did not alter any register or memory location, and that no following instruction has any dependency on the original conditional instruction. The hazard detection logic 44 may then look past the conditional instruction to one or more prior instructions, to determine a following instruction's actual dependency(s).
[0024] This process is depicted in flow diagram form in Figure 3. Following instruction decode (block 52), if an instruction is a conditional instruction (block 54), it is converted to an emissary instruction and a base instruction (block 56). The base instruction is halted at an instruction issue pipe stage 46, and the emissary instruction is issued into an available execution pipeline 12a, 12b (block 58). When the emissary instruction reaches a condition evaluation pipe stage 48, 50, it evaluates the condition (block 60) and reports the condition evaluation to the base instruction. If the condition evaluates TRUE, the base instruction is issued into an available pipeline 12a, 12b (block 62). If the condition evaluates FALSE, the base instruction is either discarded or converted to a NOP, which is issued into the pipeline 12a, 12b (block 64). In either case, and also if the decoded instruction is not a conditional instruction (block 54), hazard detection is performed for following instructions (block 66), as it is known whether or not the conditional instruction will execute.
[0025] Although the present inventive concepts have been described herein with respect to particular features, aspects and embodiments thereof, it will be apparent that numerous variations, modifications, and other embodiments are possible within the broad scope of the present inventive concepts, and accordingly, all variations, modifications and embodiments are to be regarded as being within the scope of the inventive concepts disclosed herein. The present embodiments are therefore to be construed in all aspects as illustrative and not restrictive and all changes coming within the meaning and equivalency range of the appended claims are intended to be embraced therein.

Claims

CLAIMSWhat is claimed is:
1. A method of executing a conditional instruction in a pipelined processor, the conditional instruction operative to perform an operation if a condition is satisfied, comprising: converting said conditional instruction into an emissary instruction operative to determine whether the condition is satisfied and an unconditional base instruction operative to perform said operation; executing said emissary instruction; and executing said unconditional base instruction if execution of the emissary instruction indicates that the condition is satisfied.'
2. The method of claim 1 further comprising executing a NOP instruction in place of the unconditional base instruction if execution of the emissary instruction indicates that the condition is not satisfied.
3. The method of claim 1 further comprising discarding said base instruction if execution of the emissary instruction indicates that the condition is not satisfied.
4. The method of claim 1 further comprising performing hazard detection for an instruction following said conditional instruction, after executing said emissary instruction.
5. A processor, comprising: an instruction execution pipeline including an instruction issue pipe stage and a condition evaluation pipe stage; a feedback path from said condition evaluation pipe stage to said instruction issue pipe stage; and a pipeline controller operative to convert a conditional instruction into an emissary instruction and an unconditional base instruction at said instruction issue pipe stage, issue said emissary instruction, and conditionally issue said base instruction in response to feedback from said condition evaluation pipe stage by said emissary instruction.
6. The processor of claim 5 further comprising hazard detection logic operative to detect dependencies on said conditional instruction by following instructions, after said condition evaluation.
7. A method of performing hazard detection in a pipelined processor, comprising: decoding a conditional instruction, said conditional instruction operative to perform an operation only if a condition is satisfied; converting said conditional instruction into an emissary instruction operative to evaluate said condition, and an unconditional base instruction operative to perform said operation; executing said emissary instruction to obtain a condition evaluation; and determining no following instruction is dependent on said conditional instruction if said condition is not satisfied.
8. The method of claim 7 further comprising detecting a dependency of any following instruction that reads a register or memory location written by said conditional instruction if said condition is satisfied.
9. The method of claim 8 further comprising executing said unconditional base instruction if said condition is satisfied.
10. The method of claim 9 wherein said dependency is detected prior to executing said unconditional base instruction.
11. The method of claim 10 further comprising stalling a following instruction for which a dependency is detected, until said unconditional base instruction completes execution.
12. The method of claim 7 further comprising executing a NOP instruction in place of said unconditional base instruction if said condition is not satisfied.
13. The method of claim 7 further comprising discarding said base instruction if said condition is not satisfied.
14. A method of performing hazard detection in a pipelined processor, comprising: decoding a conditional instruction where the conditional instruction is operative to perform an operation only if a condition is TRUE; converting the conditional instruction into two separate instructions, an emissary instruction, and an unconditional base instruction operative upon execution to perform the operation; executing the emissary instruction and causing the emissary instruction to evaluate the condition; executing or not executing the base instruction depending upon the evaluation of the condition by the emissary instruction; and detecting a data dependency between one or more following instructions and the base instruction.
15. The method of claim 14 further comprising, in response to detecting the dependency, delaying the execution of one or more of the following instructions until the base instruction has performed the operation.
16. The method of claim 14 further comprising, in response to detecting no dependency on the base instruction by one or more following instructions, immediately launching the one or more following instructions for execution.
17. A method of executing a conditional instruction, comprising: converting said conditional instruction into an emissary instruction and a base instruction; executing said emissary instruction, and conditionally executing said base instruction in response to a condition evaluation by said emissary instruction.
18. The method of claim 17 further comprising performing hazard detection on one or more instructions following said conditional instruction.
19. The method of claim 18 wherein no dependency exists between one or more following instructions and said conditional instruction if said base instruction is not executed.
PCT/US2006/004892 2005-02-10 2006-02-10 Conditional instruction execution via emissary instruction for condition evaluation Ceased WO2006086713A2 (en)

Priority Applications (6)

Application Number Priority Date Filing Date Title
JP2007555283A JP4903720B2 (en) 2005-02-10 2006-02-10 Conditional instruction execution via an emissary instruction for condition evaluation
DE602006013415T DE602006013415D1 (en) 2005-02-10 2006-02-10 EXECUTION OF COMMANDMENTS BY EMISSARY COMMANDS TO CONDITIONAL EVALUATION
EP06734847A EP1846817B1 (en) 2005-02-10 2006-02-10 Conditional instruction execution via emissary instruction for condition evaluation
BRPI0606973-8A BRPI0606973A2 (en) 2005-02-10 2006-02-10 conditional statement execution via messenger instruction for condition assessment
AT06734847T ATE463787T1 (en) 2005-02-10 2006-02-10 EXECUTION OF CONDITIONAL COMMANDS BY EMISSARY COMMANDS FOR CONDITIONAL EVALUATION
IL185066A IL185066A0 (en) 2005-02-10 2007-08-06 Conditional instruction execution via emissary instruction for condition evaluation

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
US11/055,919 2005-02-10
US11/055,919 US7210024B2 (en) 2005-02-10 2005-02-10 Conditional instruction execution via emissary instruction for condition evaluation

Publications (2)

Publication Number Publication Date
WO2006086713A2 true WO2006086713A2 (en) 2006-08-17
WO2006086713A3 WO2006086713A3 (en) 2006-11-16

Family

ID=36676060

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/US2006/004892 Ceased WO2006086713A2 (en) 2005-02-10 2006-02-10 Conditional instruction execution via emissary instruction for condition evaluation

Country Status (11)

Country Link
US (1) US7210024B2 (en)
EP (1) EP1846817B1 (en)
JP (1) JP4903720B2 (en)
KR (1) KR100916168B1 (en)
CN (1) CN100498693C (en)
AT (1) ATE463787T1 (en)
BR (1) BRPI0606973A2 (en)
DE (1) DE602006013415D1 (en)
IL (1) IL185066A0 (en)
TW (1) TW200641675A (en)
WO (1) WO2006086713A2 (en)

Families Citing this family (25)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2008079433A1 (en) * 2006-12-20 2008-07-03 Rowley, Richard Method and system for installing a root certificate on a computer with a root update mechanism
KR100854720B1 (en) * 2007-03-23 2008-08-27 삼성전자주식회사 Loop integration method and loop integration device
US7793079B2 (en) * 2007-06-27 2010-09-07 Qualcomm Incorporated Method and system for expanding a conditional instruction into a unconditional instruction and a select instruction
CN101739237B (en) * 2009-12-21 2013-09-18 龙芯中科技术有限公司 Device and method for realizing functional instructions of microprocessor
US8527804B2 (en) * 2010-11-01 2013-09-03 Qualcomm Incorporated Architecture and method for eliminating store buffers in a DSP/processor with multiple memory accesses
US9176733B2 (en) 2011-04-07 2015-11-03 Via Technologies, Inc. Load multiple and store multiple instructions in a microprocessor that emulates banked registers
EP2508980B1 (en) * 2011-04-07 2018-02-28 VIA Technologies, Inc. Conditional ALU instruction pre-shift-generated carry flag propagation between microinstructions in read-port limited register file microprocessor
US9292470B2 (en) 2011-04-07 2016-03-22 Via Technologies, Inc. Microprocessor that enables ARM ISA program to access 64-bit general purpose registers written by x86 ISA program
US9141389B2 (en) 2011-04-07 2015-09-22 Via Technologies, Inc. Heterogeneous ISA microprocessor with shared hardware ISA registers
US9128701B2 (en) 2011-04-07 2015-09-08 Via Technologies, Inc. Generating constant for microinstructions from modified immediate field during instruction translation
US9043580B2 (en) 2011-04-07 2015-05-26 Via Technologies, Inc. Accessing model specific registers (MSR) with different sets of distinct microinstructions for instructions of different instruction set architecture (ISA)
US9317288B2 (en) 2011-04-07 2016-04-19 Via Technologies, Inc. Multi-core microprocessor that performs x86 ISA and ARM ISA machine language program instructions by hardware translation into microinstructions executed by common execution pipeline
US9336180B2 (en) 2011-04-07 2016-05-10 Via Technologies, Inc. Microprocessor that makes 64-bit general purpose registers available in MSR address space while operating in non-64-bit mode
US9146742B2 (en) 2011-04-07 2015-09-29 Via Technologies, Inc. Heterogeneous ISA microprocessor that preserves non-ISA-specific configuration state when reset to different ISA
US9645822B2 (en) 2011-04-07 2017-05-09 Via Technologies, Inc Conditional store instructions in an out-of-order execution microprocessor
US9274795B2 (en) 2011-04-07 2016-03-01 Via Technologies, Inc. Conditional non-branch instruction prediction
US9032189B2 (en) * 2011-04-07 2015-05-12 Via Technologies, Inc. Efficient conditional ALU instruction in read-port limited register file microprocessor
US8880857B2 (en) 2011-04-07 2014-11-04 Via Technologies, Inc. Conditional ALU instruction pre-shift-generated carry flag propagation between microinstructions in read-port limited register file microprocessor
US9378019B2 (en) 2011-04-07 2016-06-28 Via Technologies, Inc. Conditional load instructions in an out-of-order execution microprocessor
US8880851B2 (en) 2011-04-07 2014-11-04 Via Technologies, Inc. Microprocessor that performs X86 ISA and arm ISA machine language program instructions by hardware translation into microinstructions executed by common execution pipeline
US8924695B2 (en) 2011-04-07 2014-12-30 Via Technologies, Inc. Conditional ALU instruction condition satisfaction propagation between microinstructions in read-port limited register file microprocessor
US9898291B2 (en) 2011-04-07 2018-02-20 Via Technologies, Inc. Microprocessor with arm and X86 instruction length decoders
US9244686B2 (en) 2011-04-07 2016-01-26 Via Technologies, Inc. Microprocessor that translates conditional load/store instructions into variable number of microinstructions
US9639359B2 (en) * 2013-05-21 2017-05-02 Advanced Micro Devices, Inc. Thermal-aware compiler for parallel instruction execution in processors
US12511122B2 (en) 2020-05-15 2025-12-30 Altera Corporation Hazard mitigation for lightweight processor cores

Family Cites Families (12)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JPS5533235A (en) * 1978-08-30 1980-03-08 Toshiba Corp Microprogram control system
US6016543A (en) * 1997-05-14 2000-01-18 Mitsubishi Denki Kabushiki Kaisha Microprocessor for controlling the conditional execution of instructions
CN100347668C (en) * 1997-08-29 2007-11-07 松下电器产业株式会社 Instruction mapping device
US6505296B2 (en) * 1997-10-13 2003-01-07 Hewlett-Packard Company Emulated branch effected by trampoline mechanism
US6170052B1 (en) * 1997-12-31 2001-01-02 Intel Corporation Method and apparatus for implementing predicated sequences in a processor with renaming
EP0992891A1 (en) * 1998-10-06 2000-04-12 Texas Instruments Inc. A processor
JP3437135B2 (en) * 1999-02-08 2003-08-18 松下電器産業株式会社 Program execution control system and method
US6598153B1 (en) 1999-12-10 2003-07-22 International Business Machines Corporation Processor and method that accelerate evaluation of pairs of condition-setting and branch instructions
US6609190B1 (en) * 2000-01-06 2003-08-19 International Business Machines Corporation Microprocessor with primary and secondary issue queue
US6877086B1 (en) 2000-11-02 2005-04-05 Intel Corporation Method and apparatus for rescheduling multiple micro-operations in a processor using a replay queue and a counter
JP3900485B2 (en) * 2002-07-29 2007-04-04 インターナショナル・ビジネス・マシーンズ・コーポレーション Optimization device, compiler program, optimization method, and recording medium
GB2402510A (en) * 2003-06-05 2004-12-08 Advanced Risc Mach Ltd Predication instruction within a data processing system

Also Published As

Publication number Publication date
DE602006013415D1 (en) 2010-05-20
EP1846817A2 (en) 2007-10-24
JP4903720B2 (en) 2012-03-28
KR100916168B1 (en) 2009-09-08
CN101151591A (en) 2008-03-26
ATE463787T1 (en) 2010-04-15
CN100498693C (en) 2009-06-10
JP2008530693A (en) 2008-08-07
US20060179288A1 (en) 2006-08-10
WO2006086713A3 (en) 2006-11-16
KR20070101393A (en) 2007-10-16
US7210024B2 (en) 2007-04-24
BRPI0606973A2 (en) 2009-07-28
TW200641675A (en) 2006-12-01
IL185066A0 (en) 2007-12-03
EP1846817B1 (en) 2010-04-07

Similar Documents

Publication Publication Date Title
US7210024B2 (en) Conditional instruction execution via emissary instruction for condition evaluation
JP5335946B2 (en) Power efficient instruction prefetch mechanism
CN101160561B (en) Suppress update of branch history register by loop ending branch
CN102934075B (en) Method and apparatus for changing the sequential flow of a program using advance notification techniques
KR100938367B1 (en) System and method for correcting branch prediction errors
US7552318B2 (en) Branch lookahead prefetch for microprocessors
KR101099203B1 (en) Methods and apparatus for emulating the branch prediction behavior of an explicit subroutine call
US7624256B2 (en) System and method wherein conditional instructions unconditionally provide output
US6728872B1 (en) Method and apparatus for verifying that instructions are pipelined in correct architectural sequence
TW201403472A (en) Optimizing register initialization operations
US6766442B1 (en) Processor and method that predict condition register-dependent conditional branch instructions utilizing a potentially stale condition register value
KR20010109354A (en) System and method for reducing write traffic in processors
CN101176060A (en) Branch target address cache storing two or more branch target addresses per index
US7444501B2 (en) Methods and apparatus for recognizing a subroutine call
US6871275B1 (en) Microprocessor having a branch predictor using speculative branch registers
US10990403B1 (en) Predicting an outcome of an instruction following a flush
HK1112087A (en) Conditional instruction execution via emissary instruction for condition evaluation
US12360769B2 (en) Branch target buffer operation with auxiliary indirect cache

Legal Events

Date Code Title Description
WWE Wipo information: entry into national phase

Ref document number: 200680009952.3

Country of ref document: CN

121 Ep: the epo has been informed by wipo that ep was designated in this application
WWE Wipo information: entry into national phase

Ref document number: 185066

Country of ref document: IL

WWE Wipo information: entry into national phase

Ref document number: 2007555283

Country of ref document: JP

NENP Non-entry into the national phase

Ref country code: DE

WWE Wipo information: entry into national phase

Ref document number: 2006734847

Country of ref document: EP

WWE Wipo information: entry into national phase

Ref document number: 1290/MUMNP/2007

Country of ref document: IN

WWE Wipo information: entry into national phase

Ref document number: 1020077020700

Country of ref document: KR

ENP Entry into the national phase

Ref document number: PI0606973

Country of ref document: BR

Kind code of ref document: A2