# US 2021/0063973 A1 · Semiconductor device and prediction method for resource usage in semiconductor device

> Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt

- Number: US 2021/0063973 A1
- Document ID: US20210063973A1
- Status: Published application
- Kind: A1
- Application number: 16/829,557
- Filed: 2020-03-25
- Published: 2021-03-04
- Earliest claimed: 2019-09-04
- Assignees: Samsung Electronics Co., Ltd.
- Applicants: Samsung Electronics Co., Ltd.
- Inventors: Jong-Lae PARK; Bum Gyu PARK; Young Tae LEE; Choong Hoon PARK
- Web page: https://patentgrep.com/patent/US20210063973A1

## CPC

- G05B 13/042: {in which a parameter or coefficient is automatically adjusted to optimise the performance}
- G05B 17/02: electric
- G06F 1/3206: Monitoring of events, devices or parameters that trigger a change in power modality
- G06F 1/324: by lowering clock frequency
- G06F 1/3296: by lowering the supply or operating voltage
- G06Q 10/06: Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
- H04L 1/203: {Details of error rate determination, e.g. BER, FER or WER}
- G05B 13/048: {using a predictor}
- Y02D 10/00: Energy efficient computing, e.g. low power processors, power management or thermal management
- G05B: CONTROL OR REGULATING SYSTEMS IN GENERAL; FUNCTIONAL ELEMENTS OF SUCH SYSTEMS; MONITORING OR TESTING ARRANGEMENTS FOR SUCH SYSTEMS OR ELEMENTS (systems for controlling or regulating non-electric variables G05D; systems for regulating electric or magnetic variables G05F; control devices or systems insofar as characterised by mechanical features only G05G)
- G06F: ELECTRIC DIGITAL DATA PROCESSING (computer systems based on specific computational models G06N)
- G06Q: INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- H04L: TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION (arrangements common to telegraphic and telephonic communication H04M)
- Y02D: CLIMATE CHANGE MITIGATION TECHNOLOGIES IN INFORMATION AND COMMUNICATION TECHNOLOGIES [ICT], I.E. INFORMATION AND COMMUNICATION TECHNOLOGIES AIMING AT THE REDUCTION OF THEIR OWN ENERGY USE

## Abstract

A semiconductor device is provided. The semiconductor device includes a processing device that provides resource usage information including a utilization value; and a prediction information generating device that generates resource usage prediction information based on the resource usage information and provides the resource usage prediction information to the processing device. The prediction information generating device includes: an error calculator to calculate an error value between the utilization value and a predicted value included in the resource usage prediction information; a margin value calculator to receive the error value from the error calculator and calculate a margin value using the error value; an anchor value calculator to calculate an anchor value using the utilization value; and a predictor to output the predicted value using the anchor value and the margin value. The processing device controls resource allocation of the processing device based on the resource usage prediction information.

## Claims

### Claim 1

A semiconductor device comprising:
a processing device configured to provide resource usage information comprising a utilization value; and
a prediction information generating device configured to generate resource usage prediction information based on the resource usage information and provide the resource usage prediction information to the processing device,
wherein the prediction information generating device comprises:
an error calculator configured to calculate an error value between the utilization value and a predicted value included in the resource usage prediction information;
a margin value calculator configured to receive the error value from the error calculator and calculate a margin value using the error value;
an anchor value calculator configured to calculate an anchor value using the utilization value; and
a predictor configured to output the predicted value by calculating a prediction constant using the anchor value and the margin value, and
wherein the processing device is further configured to control resource allocation of the processing device based on the resource usage prediction information.

### Claim 2 (depends on 1)

The semiconductor device of claim 1, wherein the error calculator comprises a subtractor, and
wherein the subtractor is configured to identify the error value as a difference between the utilization value and the predicted value.

### Claim 3 (depends on 2)

The semiconductor device of claim 2, wherein the error calculator is further configured to calculate a plurality of error values between the utilization value and the predicted value, the error value being one of the plurality of error values, and
wherein the error calculator comprises a controller configured to sequentially output the plurality of error values in a first in first out (FIFO) manner.

### Claim 4 (depends on 1)

The semiconductor device of claim 1, wherein the error calculator is further configured to identify a plurality of error values between the utilization value and the predicted value, the error value being one of the plurality of error values, and
wherein the margin value calculator is further configured to calculate the margin value using a Mean Absolute Percentage Error (MAPE) function and the plurality of error values.

### Claim 5 (depends on 1)

The semiconductor device of claim 1, wherein the anchor value calculator is further configured to calculate the anchor value using the utilization value and a Per-Entity Load Tracking (PELT) function.

### Claim 6 (depends on 1)

The semiconductor device of claim 1, wherein the predictor is further configured to identify the predicted value by adding the anchor value and the margin value.

### Claim 7 (depends on 1)

The semiconductor device of claim 1, wherein the predictor is further configured to identify the predicted value by multiplying the margin value by a weight, and adding the anchor value and the weighted margin value.

### Claim 8

A prediction method for resource usage in a semiconductor device, the prediction method comprising:
receiving, by an error calculator, a utilization value and a predicted value, and calculating an error value between the utilization value and the predicted value;
receiving, by a margin value calculator, the error value from the error calculator, and calculating a margin value using the error value;
calculating, by an anchor value calculator, an anchor value using the utilization value;
calculating, by a predictor, a prediction constant using the anchor value and the margin value and outputting the predicted value; and
controlling, by a scheduler, resource allocation of the semiconductor device based on the predicted value.

### Claim 9 (depends on 8)

The prediction method of claim 8, wherein the calculating the error value between the utilization value and the predicted value is performed by a subtractor of the error calculator.

### Claim 10 (depends on 9)

The prediction method of claim 9, wherein the calculating the error value comprises calculating a plurality of error values between the utilization value and the predicted value, and the error value is one of the plurality of error values, and
wherein the prediction method further comprises sequentially outputting the plurality of error values in a first in first out (FIFO) manner.

### Claim 11 (depends on 8)

The prediction method of claim 8, wherein the calculating the error value comprises calculating a plurality of error values between the utilization value and the predicted value, and the error value is one of the plurality of error values, and
wherein the calculating the margin value uses a Mean Absolute Percentage Error (MAPE) function and the plurality of error values.

### Claim 12 (depends on 8)

The prediction method of claim 8, wherein the calculating the anchor value comprises using the utilization value and a Per-Entity Load Tracking (PELT) function.

### Claim 13 (depends on 8)

The prediction method of claim 8, wherein the outputting the predicted value comprises adding the anchor value and the margin value.

### Claim 14 (depends on 8)

The prediction method of claim 8, wherein the outputting the predicted value comprises multiplying the margin value by a weight, and adding the anchor value and the weighted margin value.

### Claim 15

A semiconductor device comprising:
a processing device configured to provide resource usage information comprising a utilization value; and
a prediction information generating device configured to generate resource usage prediction information based on the resource usage information and provide the resource usage prediction information to the processing device,
wherein the prediction information generating device comprises:
an error calculator configured to calculate an error value using the utilization value and a predicted value included in the resource usage prediction information;
a margin value calculator connected to the error calculator;
an anchor value calculator configured to receive the utilization value; and
a predictor connected to the margin value calculator and the anchor value calculator, and configured to output the predicted value,
wherein the predicted value varies according to a difference between the utilization value and the predicted value, and
wherein the processing device is further configured to control resource allocation of the processing device based on the resource usage prediction information.

### Claim 16 (depends on 15)

The semiconductor device of claim 15, wherein the error calculator comprises a subtractor, and
wherein the subtractor is configured to identify the difference between the utilization value and the predicted value.

### Claim 17 (depends on 15)

The semiconductor device of claim 15, wherein the error calculator is further configured to calculate a plurality of error values between the utilization value and the predicted value, the error value being one of the plurality of error values, and
wherein the error calculator comprises a controller configured to sequentially output the plurality of error values in a first in first out (FIFO) manner.

### Claim 18 (depends on 15)

The semiconductor device of claim 15, wherein the error calculator is further configured to identify a plurality of error values between the utilization value and the predicted value, the error value being one of the plurality of error values,
wherein the margin value calculator is further configured to calculate a margin value using a Mean Absolute Percentage Error (MAPE) function and the plurality of error values, and
wherein the predictor is further configured to identify the predicted value using the margin value.

### Claim 19 (depends on 18)

The semiconductor device of claim 18, wherein the anchor value calculator is further configured to calculate an anchor value using the utilization value,
wherein the predictor is further configured to identify the predicted value by adding the anchor value and the margin value.

### Claim 20 (depends on 19)

The semiconductor device of claim 19, wherein the anchor value calculator is further configured to calculate the anchor value using a Per-Entity Load Tracking (PELT) function.
