Sign In |Help & Support
ALL SECTORS
  • ALL SECTORS
  • GB(National Standard)
  • CB(Shipping)
  • CECS(Engineering Construction)
  • CJ(Urban Construction)
  • CY(News and Publication)
  • DB(Provincial Standard)
  • DL(Electricity & Power)
  • DZ(Geology & Mineralogy)
  • FZ(Spinning & Textile)
  • GA(Public Security)
  • HB(Aviation)
  • HG(Chemical Industry)
  • HJ(Environmental Protection)
  • JB(Machinery)
  • JC(Building Materials)
  • JG(Building & Construction)
  • JJ(Metering)
  • JT(Highway & Transportation)
  • LY(Forestry)
  • MT(Coal)
  • NB(Energy)
  • NY(Agriculture)
  • QB(Light Industry)
  • QC(Automobile & Vehicle)
  • QJ(Aerospace)
  • SH(Petrochemical)
  • SJ(Electronics)
  • SL(Water Resources)
  • SN(Commodity Inspection)
  • SY(Oil & Gas)
  • TB(Railway & Train)
  • YB(Ferrous Metallurgy)
  • YC(Tobacco)
  • YD(Telecommunication)
  • YY(Medical Device)
Database: 365,228(8 Aug 2026)
ship engineering highway project critical-near-critical gas-oil reservoirs
GB/T 41779-2022 in English

GB/T 41779-2022 in English

VALID

Test methods of energy efficiency for high performance computer system

  • Issued on:2022-10-12
  • Implemented on:2023-05-01
  • File Format:PDF
  • Delivery:Via email within 1~3 business days
Price(USD): $220.00
$214.00
Standard No: GB/T 41779-2022
Document status: VALID
Title in English: Test methods of energy efficiency for high performance computer system
Title in Chinese: 高性能计算机系统能效测试方法
Language: English
File Format: Electronic (PDF)
Delivery: Via email within 1~3 business days
Issued on: 2022-10-12
Implemented on: 2023-05-01
ICS Classification: 35.020-Information technology (IT) in general
Chinese Classification: L04-Basic standards and general methods
Professional Classification: GB-National Standard
Related Topics: efficiency
high performance
Test method for special computer
Energy efficiency test
high efficiency photometer
high efficiency photometer
Ability Value Calculation Method
computer system
Efficient detection ability
High Energy Analysis System
High energy leak tester
Efficient Characteristic Tester
High Energy Analysis System
high energy counter
High energy ray testing system
System testing within the year
computer intelligence
high performance computer
computer system
high performance computer
Denaturation efficient
Fugao computer
Test power
Calculation method of systematic error
High Energy Particle Measurement System
System Error Calculation Method
Half Height Calculation Method
half-height calculation
Calculation method of surface energy
high performance computing
GBT41779
GB/T 41779-2022
Calculation of plant height
Can a pH meter measure organic solvents?
High energy consumption motor energy efficiency
Matrix effect calculation method
Calculation method of performance indicators of chromatographic columns
Gaussmeter measurement method
computerized measurement methods
Column efficiency test method
Mill energy efficiency
Gaussmeter magnetic testing method
First-level energy efficiency test method for water boilers
Column efficiency testing and calculation methods
Soft pack battery first efficiency calculation method
Functional liver volume calculation method
Calculation method of gauge length of tensile testing machine
Chromatography column efficiency calculation method

《GB/T 41779-2022高性能计算机系统能效测试方法》由TC28(全国信息技术标准化技术委员会)归口,TC28SC39(全国信息技术标准化技术委员会信息技术与可持续发展分会)执行,主管部门为国家标准化管理委员会。


Introduction

Interpretation of the core energy efficiency indicator system

Performance indicators Calculation method Measurement requirements
HPCEE Measured FLOPS/system energy consumption ≥5 consecutive tests
FLOPS M/(t×10^9) Triangular system solution time ≥30min

Analysis of key test technologies

Compute nodes use matrix blocking technology in the LU decomposition stage:

  • Matrix initialization formula: A = LU (dimension N×N)
  • Key calculation steps:
    A22 - L21U12 = L22U22
  • Accelerator card nodes must meet double-precision floating-point calculations

Energy efficiency classification and acceptance standards

Energy efficiency rating HPCEE value range Power consumption test error
Special grade ≥5,000 MFLOPS/kW ≤±2.5%
Level 1 2,000-4,999 MFLOPS/kW ≤±3%
Level 2 1,000-1,999 MFLOPS/kW ≤±5%

Technical Points for Implementation of Standards

  1. Environmental requirements guarantee:
    • Temperature fluctuation <±0.5℃/h
    • UPS power supply ripple factor <5%
    • See Chapter 6 Cooling subsystem adaptation specification
  2. Test tool verification:
    • Electricity monitoring equipment must achieve 0.5S-level accuracy
    • Lindpack test procedures must pass N≥1e5 large matrix verification

Inspiration for industry development

This standard unified for the first time:

  • Energy efficiency conversion method for heterogeneous computing nodes
  • Storage subsystem energy consumption calculation method

Typical application:
An AI supercomputing project passed the standard test:
The measured computing efficiency reached 73.6%, saving>820,000kWh of electricity per year.

Sample only — not a preview of GB/T 41779-2022
Page: 1 / 0
100%

Loading PDF document...

Error loading PDF. Please make sure the file is valid and try again.

We also recommend