Optimization method of microprocessor microarchitecture parameters based on simulator

A technology of microprocessor and architecture, applied in the fields of electrical digital data processing, special data processing applications, instruments, etc., can solve the problems of long time consumption, long time consumption, coarse granularity of performance statistics, etc.

Active Publication Date: 2017-09-01
NAT UNIV OF DEFENSE TECH
View PDF2 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0004] The prior art method of running a benchmark program analysis result through a clock-accurate simulator to explore the design space of a microprocessor has the following defects: first, the granularity of the performance statistics obtained by running the benchmark program on the simulator is relatively coarse, and cannot Get fine-grained execution of specific programs
For example, which piece of code takes the longest time to execute the program, what causes the time-consuming and other fine-grained information, currently cannot be obtained through the output of the simulator
Second, the design space exploration based on the coarse-grained simulation analysis results is relatively blind, and each iteration takes a long time

Method used

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View more

Image

Smart Image Click on the blue labels to locate them in the text.
Viewing Examples
Smart Image
  • Optimization method of microprocessor microarchitecture parameters based on simulator
  • Optimization method of microprocessor microarchitecture parameters based on simulator
  • Optimization method of microprocessor microarchitecture parameters based on simulator

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0069] Such as figure 1 As shown, the implementation steps of the emulator-based microprocessor microarchitecture parameter optimization method in this embodiment are as follows:

[0070] 1) Analyze the processor microarchitecture described by the clock-accurate simulator, and construct the pipeline model of the microprocessor.

[0071] see figure 2 The microarchitecture of the superscalar out-of-order pipeline shown, the pipeline includes 6 pipeline segments, namely fetch, decode, rename, issue, and execute And write back (commit), where the execution (exe) includes 2 arithmetic logic units (ALU), 1 floating point unit (FP), and 1 memory access unit (load / store). The instruction fetch segment needs to access the instruction cache, and the memory segment needs to access the data cache. To rename a segment, you need to apply for a rename buffer. The decode segment needs to read the register file. The result of the branch prediction will affect the instruction fetch. The ...

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to view more

PUM

No PUM Login to view more

Abstract

The invention discloses a simulator-based microprocessor microarchitecture parameter optimization method. The implementation steps are as follows: construct a pipeline model of the microprocessor, run the target application program through the simulator to obtain the simulation output, generate and describe the target application program in The directed acyclic graph running on the clock-accurate simulator of the target microprocessor is used as a correlation graph to generate and correct the weights of directed edges; generate critical path reports and relaxation time reports respectively according to the correlation graph; use critical path reports, Based on the slack time report, analyze the performance or power consumption bottleneck of the microprocessor, judge whether the parameters of the microarchitecture of the microprocessor need to be optimized according to the analysis results, and decide to re-execute or end the exit. The invention can quickly design and explore microarchitecture parameters such as pipeline structure and system resources in the early stage of microprocessor microarchitecture design, and has the advantages of good fine-grained optimization, good performance analysis and optimization effect, and fast and efficient optimization.

Description

technical field [0001] The invention relates to the early analysis of the performance of the microprocessor in the design stage of the microprocessor, in particular to a method for optimizing the microarchitecture parameters of the microprocessor based on a simulator. Background technique [0002] How to determine the optimal combination of microarchitecture parameters at the initial stage of design is a challenging task. "Optimal" here refers to the ability to achieve optimal performance for a certain type of application. The optimal microarchitecture parameter configuration can be achieved, and the optimal microarchitecture parameter combination should be such that there are no problems of low utilization and long waiting time of various resources in the processor when executing the target program. [0003] At present, the design space exploration is carried out in the early stage of microprocessor design. Generally, the benchmark program is run on a clock-accurate simula...

Claims

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to view more

Application Information

Patent Timeline
no application Login to view more
Patent Type & Authority Patents(China)
IPC IPC(8): G06F17/50
Inventor 王永文王蕾窦强邓宇赵天磊黄立波陈微隋兵才倪小强孙彩霞张承义高军
Owner NAT UNIV OF DEFENSE TECH
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Try Eureka
PatSnap group products