Information about http://www.energystar.gov/ia/partners/prod_development/revisions/downloads/computer/Computer_Spec_Comments_v5.0_AEA.pdf
Comments on Energy Star Program Requirements for Computers (Version 5…
Tags: austrian energy, bogner, central member, coherence, compliant products, database provision, energy agency, energy efficiency, energy star compliant, energy star program, government authorities, input and output, input and output devices, maintenance reasons, member state, national markets, procurement procedures, public procurement, purchasing office, web database,Pages: 5
Language: english
Created: Fri Mar 14 16:09:16 2008
Display cached document
Page 1
Page 2
Page 3
Page 4
Page 5
Comments on Energy Star Program Requirements
for Computers (Version 5.0 DRAFT 1)
Thomas Bogner and Bernd Schäppi, Austrian Energy Agency
Commitment
Declaration of Energy Star compliant products on lists on an annual basis:
· It would be highly recommended that new products complying with Energy Star
are to be published on the US or EU database. Provision of information on
complying products only once per year will always result in a to some extent
outdated inaccurate database which is then not useful as a source of information
for buyers.
· As a matter of fact a current EU-Regulation (EC No 106/2008) requires EU
institutions and central Member State government authorities to use energy
efficiency criteria no less demanding than those defined in the ENERGY STAR
programme when purchasing office equipment. Thus it is appreciated if
manufacturers provide data on registered products immediately after putting on
the market to support retailers and purchasers in public procurement procedures.
· We know from experience, that product information published on the EU web
database doesn't reflect actual availability in national markets. Therefore we
suggest tackling this issue in the Partner Commitment specification.
Definitions
a) General definitions for "computer"
The currently used definition for "computer" seems appropriate and sufficient still. In
aiming at more coherence the exclusion of desktop derived servers (DDS) may be
reasonable for several reasons, e.g.:
· In principle servers and of course this is valid DDS too - are directly linked
with common input and output devices only for maintenance reasons. DDS
are designed to meet completely different requirements compared to client
based products like personal computers and notebooks.
· The recently published draft for Programme Requirements for Computer
Servers serve as ideal window of opportunity thus providing a quite more
appropriate path for dealing with DDS.
b) General definitions for operational modes
Definitions for Off Mode and Sleep Mode
Definitions for off mode and sleep mode are distinct and sufficient.
Definitions for Idle State and Active State
· The proposed definitions for Idle state and Active state include some fuzziness since
active mode is an umbrella term covering Idle mode. Moreover the description of
active state as the mode in which the EEPA workload will be run offers a risk for
ambiguity.
· For the moment data for considered workloads is missing in the draft paper. We ask
for detailed information regarding the EEPA workload. In this context we recommend
to build up an empirical basis for typical and representative workloads amongst users
and stakeholders to avoid arriving at any artificial or derived workload definitions.
c) Definitions for Energy Efficiency Performance Assessment
Definitions for EEPA and EEPA Tool
· Unfortunately the explanatory notes for EEPA and EEPA Tool are rather scarce at
this point. We ask for a more comprehensive explanation of this concept enabling an
in-depth assessment.
· Especially we are interested in the tangible concept for a delineation of an evaluation
of "computer effectiveness" (line 391). Does EPA intend to define one aggregated
figure? In this case a clear specification for this concept would be needed.
Definition for Workload
· For the moment it's not clear who will define workloads independently from different
platforms (PC and MAC, versions of office suites). Will there be a feasibility for
specifying one common workload?
Energy Efficiency and Power Management Criteria
a) Power Supply Efficiency Requirements
The proposed efficiency requirements for power supplies are appreciated. Expecting further
development in this area we recommend implementing a 2-tiered approach covering still
more ambitious targets at a later stage.
b) Efficiency and Performance Requirements
Concept for Desktop, Integrated Computer, Notebook and Tablet PC Levels Energy
Efficiency performance Assessment and Associated Levels
We cherish the attempt as proposed in the draft paper to look at a more holistic view in
assessing computer efficiency.
Formula Eannual, Parameters Toff, Tsleep, Tidle, and Nactive
· We are curious about the feasibility determining a realistic annual power
consumption based on this approach. Is there an intention to a priori define values
for Toff, Tsleep and Tidle or should this be in the hand of manufactures to individually
choose their values? The same issue raises for the specification of Nactive. From our
understanding the value for Eactive will be product dependent in any case.
· In different business environments values for Ti as well for N may vary to a great
extent. Most presumably a "standardized" workload applied by the EEPA Tool will
have synthetic character and will miss everyday life relevance. Thus we doubt that
this concept based on annual energy consumption expressed in the formula
Eannual is an optimum solution. Comparison for purchasers and users may be
challenging if only one aggregated potentially misleading figure will be documented.
Common Workloads
· Our own measurements (see Annex) show that the value for On idle (in a small
tolerance band) is most significant and the outstanding factor determining the
energy consumption in operational time. In the context of real office use computing
workloads (writing texts, sending emails, doing spreadsheet analysis, using web
browsers, etc.) are negligible facing today's PC's computing performance.
· Consequently and aiming at very simple but effective approach we suggest to
define criteria for On idle mode not only for transition period but as an general
long term criterion. Moreover requirements for Psleep and Poff may also be
handled with fixed levels, as these levels are comparatively easy to handle.
· For sure the situation will be different for applications predominately run on
workstations as the related workloads will be significant in terms of energy use.
EEPA approach
· As stated in the draft document the EEPA approach shall have benefits regarding a
more effective scaling of efficiency metrics to the performance and functionality
of a given product. How can this be managed and implemented? For the moment
we are not sure about the feasibility of this approach for real world operation based
on the explanatory remarks in the document.
· We are wondering about the involvement of the EC technical group in the discussion
of the EEcoMark, since no detailed information has been made available in this
regard yet. We would appreciate comprehensive information on the current stage of
the benchmark development to be able to evaluate the concept.
Workstations levels
· Unfortunately the above mentioned approach (EEPA tool) according to the EPA
assessment would not meet the requirements for evaluation of workstations, as these
products will likely be charged with significant workloads.
Game Console Levels
· Game consoles are supposed to be challenged with distinct computing workloads.
Lacking a clear definition for idle or active mode it seems difficult setting related
requirements. Quite commonly people will use game consoles for viewing videos on
DVD too. At first glance we suggest considering "on average" levels for a replay mode
according to measurement based on IEC 62087.
Desktop derived server levels
· As mentioned above we insistently propose to cover desktop derived servers in
upcoming server specification.
Thin clients
· There are different types or designs of thin clients on the market. Thus a more
detailed differentiation of thin client types should be sought (including monitor if
applicable, number and type of interfaces, etc.)
· As a matter of fact thin clients can provide primary functionality only if they are
connected to a server. Hence the measurement of power levels should be conducted
in a network connected status requiring a distinct test procedure different to personal
computers.
Requirement regarding Qualifying Families of Products
· As stated in the draft paper in the case of multiple configurations only the highest
power configuration should be registered and belonging data reported. However for
customers seeking for most efficient products the current solution lacks information
which is relevant for purchase decision. In this regard we ask for reconsidering a
reporting requirement for the least consuming product in the specific product
family too.
Annex: Power Consumption Measurement for a Desktop PC
Measurement was accomplished with professional power analyser (sampling rate 200
kS/sec, averaging period 10 sec)
Desktop PC
100
Power [W]
80
Power [W]
60
On idle av g
40
20
0
10
20
30
40
50
00
10
30
40
50
00
10
20
20
:
:
:
:
:
:
:
:
:
:
:
:
:
:
15
15
15
15
15
16
16
16
16
16
16
17
17
17
time
Figure 1: Power Consumption over Time for a Desktop-PC
Desktop PC - load curve
100
Power [W]
80
Power [W]
60 On idle av g
40
20
0
0 100 200 300 400 500 600 700
data points
Figure 2: Load curve of Power Consumption of a Desktop-PC
Calculated share of energy consumption in active mode excluding on idle compared to
total energy consumption amount to 1,5 %, i.e. this value is below required measurement
uncertainty (according to Appendix A, Testing Requirements)