Unlock instant, AI-driven research and patent intelligence for your innovation.

A Manual Sample Set Generation Method Applied to Machine Learning Thread Division

A machine learning and sample set technology, applied in the computer field, can solve problems such as single, non-guaranteed non-regular program division, and achieve the effect of avoiding limitations

Inactive Publication Date: 2018-02-27
XI AN JIAOTONG UNIV
View PDF6 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

Its limitation is that different programs usually have different structural characteristics, and based on the traditional sample generation method to optimize all programs with a single optimization scheme, it cannot guarantee that all non-regular programs can obtain the optimal partition

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
  • A Manual Sample Set Generation Method Applied to Machine Learning Thread Division
  • A Manual Sample Set Generation Method Applied to Machine Learning Thread Division
  • A Manual Sample Set Generation Method Applied to Machine Learning Thread Division

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0051] The present invention will be described in further detail below. Said is by way of explanation and not limitation.

[0052] Generally speaking, the effect of explicit parallelization is always better than implicit parallelization. The traditional sample generation is based on implicit parallelism, while the manual sample generation of the present invention is based on explicit parallelism. Therefore, theoretically, the manual sample generation of the present invention The generation method outperforms traditional sample generation methods. The manual sample set generation process of the present invention is mainly divided into two steps: first, select a benchmark program set for TLS technical evaluation, such as Olden benchmark program set (10 programs), SPEC2006 benchmark program set (13 programs), etc.; Formally divide and explicitly tune the sp-cqip points to divide the selected benchmark programs into TLS multithreaded programs one by one (i.e., explicit paralleliz...

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 manual sample set generation method applied to machine learning thread division. First, the automatic division compiler is used to automatically divide the input application program and generate a Mips code segment with a division flag bit sp-cqip and a pre-calculated segment. , and running the code with the simulator gives the initial speedup obtained. Then manually adjust the flag and observe the change of the speedup ratio, record and save the sp‑cqip information when the speedup ratio is maximum. Finally, precomputed fragments were manually generated based on the sp‑cqip position determination. The present invention aims at the current situation that the heuristic rule division scheme uses unified division rules for all applications, and provides a sample set based on machine learning thread division schemes adapted to different application programs. The sample generation method adds a manual adjustment method on the basis of automatic generation, In this way, the potential law that the position of thread division affects the change of speedup ratio is mined.

Description

technical field [0001] The invention belongs to the technical field of computers, relates to a speculative multi-thread technology, in particular to a manual sample set generation method applied to machine learning thread division. Background technique [0002] As instruction-level parallelism encounters more and more bottlenecks and the rapid development of on-chip multiprocessors, how to use core resources more effectively has become a current research hotspot. It is speculated that multi-threading, as a kind of thread-level parallelism, has developed rapidly. Thread-level parallelism allows the existence of control and data dependencies, divides the irregular serial program into multiple thread units through the parallel compiler, and assigns them to idle processor units in turn for parallel execution, and the parallel execution of the program Correctness is determined by the underlying hardware according to the corresponding execution model. [0003] figure 1 (a), (b),...

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): G06F9/38
Inventor 赵银亮李玉祥吉烁侍加强王周恺吴强胜
Owner XI AN JIAOTONG UNIV