Results for Energy Efficiency jobs

Article
02/10/2014

Applies To: Windows 8, Windows 8.1

The results for Energy Efficiency jobs include metrics for battery duration, projected battery life, and the diagnostic metrics, in addition to the job and workload settings that were used. Depending on which workloads are run, metrics might also be collected for network traffic measured during the assessment.

Results vary depending on the mode the job runs in and the workloads you decide to use. You will always see the total duration metric (as shown in the following table) listed in the results, regardless of which mode the job runs in.

Energy Efficiency Metrics

Metric	Description
Total duration (min). This is the top-level metric for all modes.	The total time that the job runs the workloads until the end condition is met. Expand this category to see details regarding the duration of each workload in the following metrics. Duration as a percentage of all workloads that the assessment ran (Number of) Iterations

Total duration (min). This is the top-level metric for all modes.

The total time that the job runs the workloads until the end condition is met. Expand this category to see details regarding the duration of each workload in the following metrics.

Duration as a percentage of all workloads that the assessment ran
(Number of) Iterations

Battery metrics

Note

Battery life metrics in the Projected time to shutdown category are estimates and can be used to quickly identify energy efficiency issues. For the most accurate results, however, we recommend using the Total duration (min) metric, which measures the battery life from full charge to 5%.

Battery metrics are described in the following table. These metrics are gathered when you run the job in Battery run down mode or Battery run down with energy efficiency diagnostic mode.

Metric	Description
Battery count (#)	The number of batteries in the system. This is a sub-metric of Power consumed.
Capacity at last full charge	The capacity of the battery, in milliwatt hours, when it was last charged just before the job ran. This is a sub-metric of Power consumed.
Design capacity	The total capacity that the battery is designed to hold in milliwatt hours. This is a sub-metric of Power consumed.
Power at end (%)	The percentage of energy left in the battery at the end of the job. This is a sub-metric of Power consumed.
Power at end (mW h)	Amount of power remaining in the battery, in milliwatt hours, when the job ended. This is a sub-metric of Power consumed.
Power at start (%)	The percentage of charge of the battery when the job started. This is a sub-metric of Power consumed.
Power at start (mW h)	Amount of power in the battery, in milliwatt hours, when the job started. This is a sub-metric of Power consumed.
Power consumed (%)	The percentage of power consumed during the job. Expand this category to see additional details on the power consumption during this job.
Power consumed (mW h)	Amount of power, in milliwatt hours, consumed during the job.
Power consumption rate (mW h/min)	The rate of power consumption, in milliwatt hours per minute. This is a sub-metric of Projected time to shutdown (min).
Projected time to fully drain battery (min)	Projected time, in minutes, until the battery is fully drained. This is a sub-metric of Projected time to shutdown (min).
Projected time to shutdown (min)	Projected time, in minutes, from the beginning of the assessment until the computer automatically shuts down because of battery drained of energy. Expand this category to see additional details on the power consumption rate.
Time to drain 1% of battery capacity (min)	The time, in minutes, to consume one percent of battery capacity. This is a sub-metric of Projected time to shutdown (min).

Power configuration metrics

Power configuration metrics are described in the following table. These metrics are gathered when you run the job in Battery run down with energy efficiency diagnostic mode or Energy efficiency diagnostic only mode.

Metric	Description
Availability Request Count	The number of processes making availability requests (of all types).
Power Policy Configuration Errors	Power policy configuration errors is an instance component of energy efficiency issues such as device and bus low-power mode support, and misconfigured power policy settings related to quality. All tagged by line "Power Policy:" in the "Errors" section of the report.
Power Policy Configuration Warnings	Power policy configuration warnings is an instance component metric that contains the aggregated power policy settings that powercffg.exe reports as an energy efficiency issue. Such issues could include device and bus low-power mode support, and misconfigured power policy settings related to quality.
USB Selective Suspend	This is a secondary counter metric that holds the number of processes requesting a system resolution change.

CPU metrics

CPU utilization metrics are described in the following table. These metrics are gathered when you run the job in Battery run down with energy efficiency diagnostic mode or Energy efficiency diagnostic only mode.

Active C-State Residency

Active C-State Residency is captured by the highest C-State residency of the CPU, is a primary interval metric, which measure the percentage of time spend in the highest C-State.

Lowest C-State

The Lowest C-State Residency is captured by the deep C-State residency is a primary interval metric, which represents the percent of time that the CPU spend in deepest available C-State. The value of the C-States can be C2, C3, C6, and so on. This depends on the CPU architecture.

Note
A C-state is an idle state, categorized from C0, when the CPU is performing useful work, to various levels of idleness including C1, C2, and C3. In the idle states, the CPU should be conserving energy at various levels of efficiency.

Periodic CPU sources <100 ms

This metric is a secondary count metric, which counts the processes or modules that wake up the CPU more often than every 100 milliseconds.

Periodic CPU sources <300 ms

This metric is a secondary count metric, which counts the processes or modules that wake up the CPU more often for a period less than 300 milliseconds, excluding the periodic CPU sources less than 100 milliseconds.

Processes Changing System Timer resolution

This metric is a secondary counter metric which hold the number of processes requesting a system resolution change.

Processes Utilizing Greater Than 1 Percent CPU

This is a secondary counter metric, which holds the count of processes that are using more than 1 percent of the CPU Utilization.

System Timer Resolution

The system timer resolution determines how frequently windows perform two main actions: updating the timer tick count if a full tick has elapsed, and checking whether a scheduled timer object has expired.

Disk power metrics

Disk power metrics are described in the following table. These metrics are gathered when you run the job in Battery run down with energy efficiency diagnostic mode or Energy efficiency diagnostic only mode.

Metric	Description
Data read from disk	This is a primary counter metric which outputs the total size of all disk-reads in KBs, at the time of the trace subtracting the first and last 30 seconds.
Disk flush count	Disk flush count is a secondary counter metric output how many times, we received disk flushes on the hard disk.
Disk Read Count	This is a secondary counter metric which outputs the total read count of all the processes of the disk. Reading more often from the disk causes the disk cylinders to spin looking for the target track or sector, which make the disk consume more energy. The lower the score of this metric value the more energy efficient is the system.
Number of Processes by Read Count	This is a secondary counter metric which reports processes that read more than 50 times.
Number Of Processes By Read Size	This is a secondary counter metric which holds the number or processes which read more than 1 MB from the disk throughout the time period of the lifetime of the workload.
Periodic Disk Sources Less Than 10 minutes	No process or thread should write to disk frequently enough to spin it up and cause subsequent writes to the disk. This excludes writes to MFT and $logfile. This is a secondary counter metric.

Network metrics

Network metrics are described in the following table. These metrics are gathered when you run the job in any of the three available modes.

Metric	Description
Network traffic rate (KB/min)	Average rate of all the network traffic sent or received during the job run. Expand this category to see additional details on the wired and wireless network traffic.
Other traffic received rate (KB/min)	Average rate of all non-wireless network traffic received during the job run. This is a sub-metric of Network traffic rate (KB/min).
Other traffic sent rate (KB/min)	Average rate of all non-wireless network traffic sent during the job run. This is a sub-metric of Network traffic rate (KB/min).
Other traffic received (KB)	The number of all non-wireless network traffic received during the job run. This is a sub-metric of Network traffic rate (KB/min).
Other traffic sent (KB)	The number of all non-wireless network traffic sent during the job run. This is a sub-metric of Network traffic rate (KB/min).
Received network traffic rate (KB/min)	Average rate of network traffic received during the job run. This is a sub-metric of Network traffic rate.
Sent network traffic rate (KB/min)	Average rate of network traffic sent during the job run. This is a sub-metric of Network traffic rate.
Wireless received traffic rate (KB/min)	Average rate of all wireless network traffic received during the job. This is a sub-metric of Network traffic rate (KB/min).
Wireless sent traffic rate (KB/min)	Average rate of all wireless network traffic sent during the job. This is a sub-metric of Network traffic rate (KB/min).
Wireless received traffic (KB)	The number of all wireless network traffic received during the job. This is a sub-metric of Network traffic rate (KB/min).
Wireless sent traffic (KB)	The number of all wireless network traffic sent during the job. This is a sub-metric of Network traffic rate (KB/min).

Issues

The Energy Efficiency job shows the analysis that the job has produced, according to the settings that you selected. It identifies problems, provides some background on how these issues affect your system, and suggests possible remediation steps. Issues are surfaced either from the power CFG analysis or from the analysis of the ETL trace files generated by the job.

The Energy Efficiency job does not generate any issues when it is run in the Battery run down only mode. Idle efficiency issues only appear if you have added the Idle energy efficiency workload or run other workloads that include a minimum of at least 10 minutes of idle time. This is because all issues that are generated are derived from idle periods during workloads. However, if the idle period of any workload does not last longer than 10 minutes, no issues are generated for that workload.

The following issues and recommendations can appear after you run the Energy Efficiency job in either the Battery run down with energy efficiency diagnostic mode or the Energy Efficiency with diagnostic only mode.

Issue	Description	Recommendation
Highest C-State Residency	CPU utilization during idle was a certain percentage, which is above the target of 1 percent for a particular idle phase. High CPU utilization during idle can prevent the CPU from entering lower power states, reducing idle efficiency.	Ensure that there are no processes using greater than 1 percent of the CPU during idle.
C-State Residency	The CPU was in the lowest available power state only a certain percentage of the time, which is below the target percentage for an idle phase. High CPU utilization and short wake intervals can reduce this metric, which causes much poorer idle efficiency.	Investigate process CPU utilization and average wake intervals.
System Timer Resolution	The system timer resolution was a certain number of milliseconds, which is below the default maximum value of 15.6 milliseconds. This can reduce the CPU’s residency in low power states.	Investigate processes that change the system timer resolution.
Disk Reads	The disk was read a certain number of times during idle, which is above the target number of reads. Each read causes the disk to spin up, reducing overall idle efficiency.	Investigate processes reading from disk during idle.
Data Read From Disk (MB)	A number of megabytes was read from disk during idle, which is above the target read number. Reading large amounts of data causes the disk to spin up for longer periods of time, reducing overall idle efficiency.	Investigate processes reading from disk during idle.
Disk Flushes	The disk was flushed a certain number of times during idle. Each flush causes the disk to spin up, reducing overall idle efficiency.	Investigate processes that write to disk more often than once every ten minutes.
USB Device - USB Selective Suspend	A particular USB device did not enter the suspend state. Processor power management may be prevented if this device does not enter the suspend state when the device is not in use.	Update the driver or remove the device.
Display Availability Request	The driver, process, service, or program made a display availability request, which prevents the system from automatically turning off the display.	Update or remove the software generating the request, or prevent name from placing the request by using the powercfg /requestsoverride command.
A certain driver, process, or service made a system availability request	The driver, or process, or service made a system availability request, which prevents the system from entering a low-power state.	Update or remove the software that generates the request, or prevent the software from placing the request by using the powercfg /requestsoverride command.
A driver, process, or service made an away mode availability request	The driver, process, or service made an away mode availability request, which prevents the system from entering a low-power state.	Update or remove the software that generates the request, or prevent the software from placing the request by using the powercfg /requestsoverride command.
Process.exe - CPU Utilization	Process.exe, a process identifier (pid), had a certain percentage of CPU utilization during idle, which is greater than the target of 1 percent. This will increase overall CPU utilization, which prevents the CPU from entering a lower power state.	If you are not using this software, consider removing it. Otherwise, try updating the software or exploring the issue further by using Windows® Performance Analyzer (WPA).
Process.exe - Timer Interrupts	Process.exe (pid) contains timers that expire at an average interval of less than [300ms \| 100ms]. These timers may cause the CPU to wake more often, potentially reducing its average idle interval and increasing power consumption.	If you are not using this software, consider removing it. Otherwise, try updating the software or exploring the issue further by using WPA.
Process.exe - Timer Resolution	Process.exe (pid) requested an increase in system timer resolution. This can cause the CPU to wake up more frequently, reducing idle efficiency.	If you are not using this software, consider removing it. Otherwise, try updating the software.
Process.exe - Disk Utilization	Process.exe (pid) wrote to disk at an average interval of less than 10 minutes, which causes the disk to spin up and lowering idle efficiency.	If you are not using this software, consider removing it. Otherwise, try updating the software or exploring the issue further by using WPA.
Process.exe - Disk Read Count	Process.exe (pid) read from disk a certain number of times, which is greater than the target of 50 reads. Each of these reads causes the disk to spin up, lowering idle efficiency.	If you are not using this software, consider removing it. Otherwise, try updating the software or exploring the issue further by using WPA.
Process.exe - Disk Read Size	Process.exe read a certain number of megabytes from disk, which is greater than the target of 1 megabyte. Large reads extend the amount of time the disk must operate, which causes it to consume more power and reduce idle efficiency.	If you are not using this software, consider removing it. Otherwise, try updating the software or exploring the issue further by using WPA.
Driver Interrupts	A certain driver contains device interrupts that occur at an average interval of a number of milliseconds, which is less than the target number of milliseconds. These timers may cause the CPU to wake more often, reducing its average time in the lower C-State and increasing power consumption.	If you are not using this software, consider removing it. Otherwise, try updating the software.
Disk Flushes	The disk was flushed a certain number of times during idle. Each flush causes the disk to spin up, reducing idle efficiency.	Investigate processes issuing disk flushes.
Disk Read KB	A certain process (.exe) read a quantity of kilobytes from the disk, which is greater than the target quantity. Large reads extend the amount of time the disk must operate, which causes it to consume more power and reducing idle efficiency.	If you are not using this software, consider removing it. Otherwise, try updating the software.
Timer Interrupts	A driver contains device interrupts that occur at an average interval of a certain number of milliseconds, which is less than the target number of milliseconds. These timers may cause the CPU to wake more often, reducing its average time in the lower C-State and increasing power consumption.	If you are not using this software, consider removing it. Otherwise, try updating the software.

This information may include links to view associated reports or trace-log timelines.