I.
Desktop, Integrated Desktop, Notebook Computers: Below is a sample TEC calculation intended to show how levels for compliance are determined based on functional adders and operational mode measurements.
Following is a sample ETEC evaluation for a 2,0 GHz, dual core Notebook with Switchable Graphics, 8 GB Memory, Energy Efficient Ethernet (EEE), and 1 hard disk drive (HDD).
A. Measure values using the ENERGY STAR Computers Test Method:
Off Mode = 1,0 W
Sleep Mode = 1,7 W
Long Idle State = 8,0 W
Short Idle State = 10,0 W
B. Determine the proxy support provided by the operating system and network card. This is a manufacturer-reported parameter.
On Mac computers, “Wake for network access” enabled within the Energy Saver/Power Adapter Preferences signifies Base Capability or better.
On Windows computers, “ARP Offload” or “NS Offload” or similar enabled within the Advanced Properties of the Network Interface Card (accessed through the Device Manager) signifies Base Capability or better. OEM can provide further guidance on how to confirm Proxy Support
C. Calculate ETEC from power measurements and mode weightings—this example assumes no Proxy Support/Conventional Weightings:
TOFF | 25 %
TSLEEP | 35 %
TLONG_IDLE | 10 %
TSHORT_IDLE | 30 %
ETEC
8760
1000
POFFTOFFPSLEEPTSLEEPPLONG_IDLETLONG_IDLEPSHORT_IDLETSHORT_IDLE
ETEC
8760
1000
1,0 W25 %1,7 W35 %8,0 W10 %10,0 W30 %
ETEC = 40,7 kWh/year
D. Determine which Base TEC allowance applies based on graphics capability and performance score: P = [# of CPU cores] × [CPU clock speed (GHz)] = 2×2 GHz = 4.
Category Name | Graphics Capability | Notebook
Performance Score, P | Base Allowance
I1 | Integrated or Switchable Graphics | 2 < P ≤ 5,2 | 22,0
E. Determine which Functional Adder Allowances apply:
Memory: 8 GB installed, so a TECMEMORY allowance of 8 GB0,8 kWhGB6,4 kWh applies
Discrete Graphics? No, therefore TECGRAPHICS allowance does not apply.
Switchable Graphics? Yes, but TECSWITCHABLE allowance does not apply to Notebooks.
Energy Efficient Ethernet (EEE)? Yes, and assuming one EEE-compliant Ethernet port, a TECEEE allowance of 8,76 × 0,2 × (0,10 + 0,30) = 0,7 kWh applies
Storage? No, the notebook has only one hard disk drive, so no storage allowance applies.
Integrated Display? Yes, and assuming a non-enhanced performance, 14 inch display with an area of 83,4 square inches and a resolution of 1,05 megapixels, a TECINT_DISPLAY allowance of 8,76 × 0,30 × (1 + EP) × (2 × r + 0,02×A) = 8,76 × 0,30 × (2 × 1,05 MP + 0,02×83,4 in2) = 9,9 kWh applies.
F. Calculate ETEC_MAX:
ETEC_MAX = 22,0 kWh + 6,4 kWh + 0,7 kWh + 9,9 kWh
ETEC_MAX = 39,0 kWh/yr
G. Compare ETEC to the ETEC_MAX to determine if the model qualifies:
40,7 kWh/yr > 39,0 kWh/yr
Therefore, the Notebook does not meet ENERGY STAR requirements.
II.
Workstations: Below is a sample PTEC calculation for a Workstation with 2 hard drives and no Energy Efficient Ethernet capability.
A. Measure values using the ENERGY STAR Computers Test Method:
Off Mode = 2 W
Sleep Mode = 4 W
Long Idle State = 50 W
Short Idle State = 80 W
Max Power = 180 W
B. Note number of Hard Drives installed: Two hard drives installed during test.
C. Calculate PTEC from power measurements and mode weightings using Equation 4:
TOFF | TSLEEP | TLONG_IDLE | TSHORT_IDLE
35 % | 10 % | 15 % | 40 %
PTEC = (35 % × POFF + 10 % × PSLEEP + 15 % × PLONG_IDLE + 40 % × PSHORT_IDLE)
PTEC = (35 % × 2 W + 10 % × 4 W + 15 % × 50 W + 40 % × 80 W)
PTEC = 40,6 W
D. Calculate the PTEC_MAX requirement using Equation 5:
PTEC_MAX = 0,28× (PMAX + NHDD × 5) + 8,76×PEEE×(TSLEEP + TLONG_IDLE + TSHORT_IDLE)
PTEC_MAX = 0,28×(180 + 2 × 5) + 8,76 × 0 × (TSLEEP + TLONG_IDLE + TSHORT_IDLE)
PTEC_MAX = 53,2 + 0
E. Compare PTEC to the ENERGY STAR levels to determine if the model qualifies:
40,6 W ≤ 53,2 W
Therefore, the Workstation meets ENERGY STAR requirements.
The following test method shall be used for determining product compliance with requirements in the ENERGY STAR Specification for Computers.
ENERGY STAR test requirements are dependent upon the feature set of the product under evaluation. The following guidelines shall be used to determine the applicability of each section of this document:
The procedure in Section 6 shall be conducted on all eligible products that are covered under the scope as defined in Section 2 of the ENERGY STAR Final Draft Eligibility Criteria for Computers.
The procedure in Section 7 shall be conducted only on eligible Workstation Computer products.
Unless otherwise specified, all terms used in this document are consistent with the definitions in the ENERGY STAR Specification for Computers.
Test setup and instrumentation for all portions of this procedure shall be in accordance with the requirements of European Standard EN 50564:2011 (derived from IEC 62301:2011) “Electrical and electronic household and office equipment — Measurement of low power consumption”, Section 4, “General Conditions for Measurements”, unless otherwise noted in this document. In the event of conflicting requirements, the ENERGY STAR test method shall take precedence.
A. Input Power: Products intended to be powered from alternating current (ac) mains shall be connected to a voltage source appropriate for the intended market, as specified in Table 13 and Table 14.
Market | Voltage | Voltage Tolerance | Maximum Total Harmonic Distortion | Frequency | Frequency Tolerance
Europe, Australia, New Zealand | 230 V ac | +/– 1,0 % | 2,0 % | 50 Hz | +/– 1,0 %
Market | Voltage | Voltage Tolerance | Maximum Total Harmonic Distortion | Frequency | Frequency Tolerance
Europe, Australia, New Zealand | 230 V ac | +/– 4,0 % | 5,0 % | 50 Hz | +/– 1,0 %
B. Ambient Temperature: Ambient temperature shall remain between 18 °C and 28 °C, inclusive, for the duration of the test.
C. Relative Humidity: Relative humidity shall remain between 10 % and 80 %, inclusive, for the duration of the test.
D. Light Measuring Device (LMD): All LMDs shall meet the following specifications:
Accuracy: ± 2 % (± 2 digits) of the digitally displayed value; and
Acceptance Angle: 3 degrees or less.
The overall tolerance of LMDs is found by taking the absolute sum of 2 %of the targeted screen luminance and a 2 digit tolerance of the displayed value's least significant digit. For example, if the screen luminance value is 90 candela per meter squared (cd/m2) and the LMD's least significant digit is a tenth of one cd/m2, 2 % of 90 cd/m2 would be 1,8 cd/m2 and a 2 digit tolerance of the least significant digit would be 0,2 cd/m2. Thus, the displayed value would need to be 90 ± 2 cd/m2 (1,8 cd/m2 + 0,2 cd/m2).
Note: The term “nit” is sometimes used instead of the official SI unit cd/m2. One nit is equivalent to one cd/m2.
E. Power Meter: Power meters shall possess the following attributes:
Crest Factor:
An available current crest factor of 3 or more at its rated range value; and
A bound on the current range of 10 milliamperes (mA) or less.
Minimum Frequency Response: 3,0 kilo-hertz (kHz)
Minimum Resolution:
0,01 W for measurement values less than 10 W;
0,1 W for measurement values from 10 W to 100 W; and
1,0 W for measurement values greater than 100 W.
Measurement Accuracy: Measurement uncertainty as introduced by the instrument that measures the input power to the unit under test (UUT), including any external shunts.
Power measurements with a value greater than or equal to 0,5 W shall be made with an uncertainty of less than or equal to 2 % at the 95 % confidence level.
Power measurements with a value less than 0,5 W shall be made with an uncertainty of less than or equal to 0,01 W at the 95 % confidence level.
The Test Conduct shall be carried out according to the requirements in European Standard EN 62623:2013 (identical to IEC 62623:2012) “Desktop and Notebook Computers — Measurement of Energy Consumption” reference with the following guidance.
A. Small-Scale Servers, Thin Clients, and Workstations shall be configured in a manner identical to Desktops (non-integrated) unless otherwise specified. Slates/Tablets shall be configured in a manner identical to Notebooks unless otherwise specified. Portable All-In-One Computers shall be configured in a manner identical to Integrated Desktops unless otherwise specified.
Thin Clients shall run intended terminal/remote connection software during all tests.
B. Wake on LAN (WoL) settings shall be in as shipped condition for testing Sleep Mode and Off Mode.
C. For models that do not offer a Sleep Mode enabled by default, Section 6.2 shall measure power in the lowest-latency user-activated mode or state that preserves machine state and is enabled by default.
If no such state separate from Long Idle State or Off Mode exists, the measurement in Section 6.2 shall be skipped.
D. For Long Idle Mode Testing (Section 6.3), the UUT shall be allowed no more than 20 minutes from the point of ceased user input before measurements must be started. If any default settings cause the UUT to enter Long Idle after 20 minutes, begin taking measurements when the UUT has reached the 20 minute mark. Display sleep settings shall be set to default for Long Idle Mode Testing.
E. For Short Idle Mode Testing (Section 6.4), the UUT shall be allowed no more than five minutes from the point of ceased user input before measurements must be taken. Display sleep settings shall be disabled for Short Idle Mode Testing. If any other default settings cause the UUT to exit Short Idle during the measurement time, extend the settings so that the UUT remains in short idle for the duration of the measurement.
F. Desktops, Integrated Desktops, Notebook Computers, Portable All-In-One Computers, and Slates/Tablets shall be tested for Idle, Sleep, and Off Mode with Full Network Connectivity (“Proxying”) features using the as shipped setting.
G. Cellular network connections shall be disabled for testing. Additionally, Bluetooth should be left as-shipped.
A. Before performing any tests, disable display dimming, display Sleep Mode, Computer Sleep Mode, and automatic brightness control (ABC) in the Computer settings. Document all settings that were changed from the default configuration.
If ABC cannot be disabled, position a light source such that at least 300 lux directly enters the ABC sensor.
B. Display the three vertical bar video signal as defined in section 3.2.1.3 of European Standard EN 60107-1:1997 (identical to IEC 60107-1:1997) “Methods of measurement on receivers for television broadcast transmissions — Part 1: General conditions — Measurements at radio and video frequencies”. The three bar image shall be configured using the default image display application.
C. Devices with a cold cathode fluorescent lamp (CCFL) backlight shall warm-up for at least 30 minutes. All other displays shall warm-up for at least 5 minutes.
D. With the LMD, measure the luminance in the centre of the display.
E. Calibrate the UUT display brightness to the closest brightness setting that is at least 90 cd/m2 for Notebook Computers, at least 150 cd/m2 for Integrated Desktop Computers, Portable All-In-One Computers and Slates/Tablets. If the UUT's brightest setting cannot achieve the specified brightness, then set the UUT display to the brightest setting.
F. the display shall be configured with the ENERGY STAR test image, which can be found in https://www.energystar.gov/ia/partners/images/ComputerTestingImage.bmp. For Desktops, Integrated Desktops, Notebook Computers and Portable All-In-One Computers it may be set as the “desktop background” (wallpaper) or shown via an image display application. The image shall be scaled to completely fill the display area. For Slates/Tablets, the display shall be configured with the default image display application.
G. For all testing specified in Section 6, the UUT shall not be rebooted or restarted until after the power measurements for Long Idle Mode and Short Idle Mode tests are taken.
H. Slate/Tablet and Portable All-In-One Computers shall be tested with a docking station only if it is shipped with the product and is the only way to power the device mains.
UUT preparation shall be performed according to European Standard EN 62623:2013 (identical to IEC 62623:2012), Section 5.2: Test Setup; with the additional guidance in Section 5 of this document.
Sleep Mode power shall be measured according to European Standard EN 62623:2013 (identical to IEC 62623:2012), Section 5.3.3: Measuring Sleep Mode; with the additional guidance in Section 5 of this document.
Long Idle Mode power shall be measured according to European Standard EN 62623:2013 (identical to IEC 62623:2012), Section 5.3.4: Measuring Long Idle Mode; with the additional guidance in Section 5 of this document.
Short Idle Mode power shall be measured according to European Standard EN 62623:2013 (identical to IEC 62623:2012), Section 5.3.5: Measuring Short Idle Mode; with the additional guidance in Section 5 of this document.
Off Mode power shall be measured according to European Standard EN 62623:2013 (identical to IEC 62623:2012), Section 5.3.2: Measuring Off Mode; with the additional guidance in Section 5 of this document.
For Notebook Computers, repeat the Short Idle test with the display brightness set to the closest setting that is at least 150 cd/m2.
The maximum power for Workstations is found by the simultaneous operation of two industry standard benchmarks: Linpack to stress the core system (e.g., processor, memory, etc.) and SPECviewperf® (latest available version for the UUT) to stress the system's Graphics Processing Unit (GPU). This test shall be repeated three times on the same UUT, and all three measurements shall fall within a ± 2 % tolerance relative to the average of the three measured maximum power values. The average power should be used for qualification and/or TEC calculations.
Additional information on these benchmarks, including free downloads, can be found at the following locations as specified in Table 15.
Benchmark | Website
Linpack | http://www.netlib.org/linpack/
SPECviewperf | http://www.spec.org/benchmarks.html#gpc
A. UUT Preparation:
Connect a power meter capable of measuring true power to an AC line voltage source set to the appropriate voltage/frequency combination for the test. The meter shall have all the attributes listed in Section 4.1 E). The meter shall also store and output the maximum power measurement reached during the test or be capable of another method of determining maximum power.
Plug the UUT into the measurement power outlet on the meter. No power strips or uninterruptible power supply (UPS) units shall be connected between the meter and the UUT.
Record the ac voltage.
Boot the UUT and, if not already installed, install Linpack and SPECviewperf as indicated on the above Websites.
Set Linpack with all the defaults for the given architecture of the UUT and set the appropriate array size “n” for maximizing power draw during the test.
Ensure all technical guidelines relevant to running the benchmark set by the Standard Performance Evaluation Corporation (SPEC) organization for running SPECviewperf have been met.
For additional information regarding Linpack setup, see Section 9.1 Typical Linpack Starting Parameters.
B. Maximum Power Testing:
Set the meter to begin accumulating true power values at a rate greater than or equal to one reading per second, and begin taking measurements.
Run SPECviewperf and as many simultaneous instances of Linpack as needed to fully stress the system. Recommended setup information can be found in Section 9.1 C).
Accumulate power values until SPECviewperf and all Linpack instances have completed running. Record the maximum power value attained during the test.
The following data shall also be recorded:
Value of “n” (the array size) used for Linpack;
Number of simultaneous copies of Linpack run during the test;
Version of SPECviewperf run for test;
All compiler optimizations used in compiling Linpack and SPECviewperf; and
A precompiled binary for end users to download and run both SPECviewperf and Linpack. These can be distributed either through a centralized standards body such as SPEC, by the original equipment manufacturer (OEM), or by a related third party.
The benchmark test shall be performed by running both benchmarks listed below separately. The UUT shall be rebooted before testing with each benchmark. Additional information on these benchmarks, including downloads, can be found at the following locations specified in Table 16. All testing shall be performed with the latest available version of the benchmarks.
Benchmark | Website
Linpack | http://www.netlib.org/linpack/
SPECviewperf | http://www.spec.org/benchmarks.html#gpc
A. UUT Preparation:
The UUT shall be setup identical to Step 1) through Step 4) of Section 7.1 A)
If not already installed, install the benchmark as indicated on the websites listed in Table 16.
Configure the benchmark as specified in Section 7.2 B).
Time Measurement: Time measurements may be performed with a standard stopwatch or other time keeping device with a resolution of at least 1 second.
B. Benchmark Configurations:
Linpack
Configure the Linpack settings identically to the maximum power workstation test (e.g. Follow Step 5) and Step 7) of Section 7.1 A)).
Run as many simultaneous instances of Linpack as needed to fully stress the system. Recommended settings would be to set the number of simultaneous instances of Linpack equal to the number of logical and/or physical CPU cores of the system.
SPECviewperf
Configure the settings identically to the maximum power workstation test (e.g. Follow Step 6) of Section 7.1 A)).
C. Benchmark Testing:
Set the meter to begin accumulating true power values at a rate of greater than or equal to one reading per second and begin power and time measurement.
Execute the benchmark.
Stop time measurement and accumulate power values for the entire duration of the benchmark run.
The following data shall be reported:
Linpack
Value of “n” (the array size) used for Linpack;
Number of instances of Linpack simultaneously run on the system;
All compiler options used in compiling Linpack;
Energy consumed over the duration of the test; and
Linpack output file in text format which contains system performance in floating point operations per second (Flops) in addition to other Linpack parameters (e.g. number of tests, problem size, etc.).
SPECviewperf
Version of SPECviewperf used;
All compiler optimizations used in compiling SPECviewperf;
Duration of the test;
Energy consumed over the duration of the test; and
All files and folders present in the Result folder of SPECviewperf suite shall be reported.
A. European Standard EN 50564:2011 (derived from IEC 62301:2011), Electrical and electronic household and office equipment — Measurement of low power consumption.
B. European Standard EN 60107-1:1997 (identical to IEC 60107-1:1997), Methods of measurement on receivers for television broadcast transmissions — Part 1: General Considerations — Measurements at radio and video frequencies.
C. European Standard EN 62623:2013 (identical to IEC 62623:2012), Desktop and notebook computers — Measurement of energy consumption
Below are some typical starting values for the use of Linpack for testing Workstations. These values are starting points and not meant to be binding. The tester is free to use the settings most advantageous to their UUT. Platform and Operating System (OS) will have a significant impact on the applicability of these starting values. The below assumes Linux as the test OS.
A. Number of equations (problem size): See Equation.
B. Leading dimensions of array: See Equation.
The matrix size (the combination of number of equations and leading dimensions of array) should be the maximum size that will fit in the Random Access Memory (RAM) on the machine. This AWK script will calculate matrix size on a Linux machine:
awk '
BEGIN {
printf ”Maximum matrix dimension that will fit in RAM on this machine:”
}
/^MemTotal:/{
print int(sqrt(($2*1000)/8)/1000) ”K”
}
'/proc/meminfo
Use the output of this to determine what matrix size to input for both the “Number of equations” and “Leading dimensions of array” inputs. The “Number of equations” will be equal to the printed output. The “Leading dimensions of the array” will be the output rounded up to the nearest multiple of eight.
This calculation can be most easily calculated by taking the memory size, in bytes, of the UUT (denoted as m) and substituting m in Equation 1.
m1000
8
1000
Equation 9: Memory Size Calculation
C. Number of trials: c – 1 where c equals the number of logical and/or physical CPU cores of the system. The tester needs to determine which is more advantageous for the unit. The – 1 leaves one core open for use by SPECviewperf.
D. Data alignment value: Typically four with Linux systems. The best value to use is the page size boundary of the OS.