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

On-chip optical network hotspot prediction method based on LSTM neural network

A neural network and network hotspot technology, which is applied in the field of LSTM-based on-chip optical network hotspot prediction, can solve problems such as shortened chip life, attacking the system, and third-party personnel are not completely credible, and achieve good prediction results

Pending Publication Date: 2022-06-24
CHONGQING UNIV OF POSTS & TELECOMM
View PDF0 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

For chips, the existence of hotspots will increase the temperature of nodes and increase power consumption, which will greatly shorten the life of chips. To solve this problem, researchers usually use the method of estimating hotspots to balance the communication traffic of each node.
In addition, in order to shorten the time and save costs, the chip will be in contact with third-party personnel many times during the design and manufacturing process, but these third-party personnel are not completely credible, because they may implant hardware Trojan horses ( Hardware Trojan, HT) to attack the entire system

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
  • On-chip optical network hotspot prediction method based on LSTM neural network
  • On-chip optical network hotspot prediction method based on LSTM neural network
  • On-chip optical network hotspot prediction method based on LSTM neural network

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0050] In this embodiment, an 8×8 Mesh on-chip optical network topology is selected, and the model is built based on Keras, an advanced API in Python. The Loss, MAE and RMSE of the three models under different training rounds are as follows: image 3 , 4 and 5, and the comparison of the training set R2 of the three models under different training rounds is shown in Table 1.

Embodiment 2

[0052] In this embodiment, a 4×4 Mesh on-chip optical network topology is selected, and the model is built based on Keras, an advanced API in Python. The Loss of the three models under different training rounds is as follows: image 3 The R2 comparisons of the three models under different training rounds are shown in Table 2.

Embodiment 3

[0054] In this embodiment, a 6×6 Mesh on-chip optical network topology is selected, and the model is built based on Keras, an advanced API in Python. The Loss of the three models under different training rounds is as follows: image 3 The R2 comparisons of the three models under different training rounds are shown in Table 3.

[0055] It can be seen that the embodiments of the present invention all show good performance in on-chip optical networks with different core counts.

[0056] Through the above prediction process, it is possible to obtain the traffic prediction distribution diagrams of the on-chip optical network at different times in the future under different core numbers of the on-chip optical network under different embodiments. Figure 7 , 8 and 9 shown. For the determination of hot spots, the traffic value corresponding to the node is divided into n intervals, the proportion of nodes in each traffic interval is calculated, and 10% of the interval with larger tr...

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 requests to protect an on-chip optical network hotspot prediction method based on an LSTM neural network. The method comprises the following steps: cleaning, normalizing and dividing a traffic value of each node in an on-chip optical network to obtain training data and test data; a multi-input and multi-output LSTM neural network model is built for adapting to the multi-node feature in the on-chip optical network, and training data is input into the model for training; and after obtaining a training model, inputting data into the model to obtain a predicted flow value of each node. Compared with a traditional modeling method, the LSTM neural network has the advantages of being high in self-learning and self-adaption and the like, and therefore the hot spot change conditions of the nodes in the network can be analyzed and predicted through the characteristics. Compared with a current typical prediction model recurrent neural network (RNN) and a gate recurrent unit (GRU), the mean square error of the model is reduced by 8.57% and 15.7% respectively, and the fitting degrees are improved by 3.35% and 1.73% respectively.

Description

technical field [0001] The invention relates to communication technology, in particular to an LSTM-based on-chip optical network hotspot prediction method. Background technique [0002] At this stage, high-performance computing has a large number of needs and applications in numerical simulation, life science and large-scale engineering computing, and on-chip multi-core systems play a key role. Optical Network on-Chip (ONoC) is an "on-chip communication" solution to solve the problem of data transmission between different cores in an on-chip multi-core system. , SoC) and the shortcomings of poor reliability and high energy consumption in the Network on Chip (NoC), the ability of multi-core chips to process data has also been continuously improved after the on-chip optical network was proposed. At the same time, in order to improve network performance, researchers usually divide upper-layer applications into multiple tasks and map them to several adjacent intellectual proper...

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 Applications(China)
IPC IPC(8): G06N3/04G06N3/067G06N3/08G06E3/00G06F15/78
CPCG06N3/067G06N3/08G06F15/7825G06E3/006G06N3/044
Inventor 仇星郭鹏星侯维刚何香玉
Owner CHONGQING UNIV OF POSTS & TELECOMM