Re: [dpdk-dev] [PATCH v8 3/4] doc/guides/proguide/power-man: update the power API

["Hunt, David" <david.hunt@intel.com>]

Hi Liang,


On 17/9/2018 2:30 PM, Liang Ma wrote:
> Update the document for empty poll API.
>
> Signed-off-by: Liang Ma <liang.j.ma@intel.com>
> ---
>   doc/guides/prog_guide/power_man.rst | 90 +++++++++++++++++++++++++++++++++++++
>   1 file changed, 90 insertions(+)
>
> diff --git a/doc/guides/prog_guide/power_man.rst b/doc/guides/prog_guide/power_man.rst
> index eba1cc6..056cb12 100644
> --- a/doc/guides/prog_guide/power_man.rst
> +++ b/doc/guides/prog_guide/power_man.rst
> @@ -106,6 +106,96 @@ User Cases
>   
>   The power management mechanism is used to save power when performing L3 forwarding.
>   
> +
> +Empty Poll API
> +--------------
> +
> +Abstract
> +~~~~~~~~
> +
> +For packet processing workloads such as DPDK polling is continuous.
> +This means CPU cores always show 100% busy independent of how much work
> +those cores are doing. It is critical to accurately determine how busy
> +a core is hugely important for the following reasons:
> +
> +        * No indication of overload conditions
> +        * User do not know how much real load is on a system meaning
> +          resulted in wasted energy as no power management is utilized
> +
> +Compared to the original l3fwd-power design, instead of going to sleep
> +after detecting an empty poll, the new mechanism just lowers the core frequency.
> +As a result, the application does not stop polling the device, which leads
> +to improved handling of bursts of traffic.
> +
> +When the system become busy, the empty poll mechanism can also increase the core
> +frequency (including turbo) to do best effort for intensive traffic. This gives
> +us more flexible and balanced traffic awareness over the standard l3fwd-power
> +application.
> +
> +
> +Proposed Solution
> +~~~~~~~~~~~~~~~~~
> +The proposed solution focuses on how many times empty polls are executed.
> +The less the number of empty polls, means current core is busy with processing
> +workload, therefore, the higher frequency is needed. The high empty poll number
> +indicates the current core not doing any real work therefore, we can lower the
> +frequency to safe power.
> +
> +In the current implementation, each core has 1 empty-poll counter which assume
> +1 core is dedicated to 1 queue. This will need to be expanded in the future to
> +support multiple queues per core.
> +
> +Power state definition:
> +^^^^^^^^^^^^^^^^^^^^^^^
> +
> +* LOW:  Not currently used, reserved for future use.
> +
> +* MED:  the frequency is used to process modest traffic workload.
> +
> +* HIGH: the frequency is used to process busy traffic workload.
> +
> +There are two phases to establish the power management system:
> +^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
> +* Initialization/Training phase. The training phase is necessary
> +  in order to figure out the system polling baseline numbers from
> +  idle to busy. The highest poll count will be during idle, where all
> +  polls are empty. These poll counts will be different between
> +  systems due to the many possible processor micro-arch, cache
> +  and device configurations, hence the training phase.
> +  In the training phase, traffic is blocked so the training algorithm
> +  can average the empty-poll numbers for the LOW, MED and
> +  HIGH  power states in order to create a baseline.
> +  The core's counter are collected every 10ms, and the Training
> +  phase will take 2 seconds.
> +  Training is disabled as default configuration.
> +  The default parameter is applied. Simple App still can trigger
> +  training if that's needed

Suggest:

* Training phase. This phase is used to measure the optimal frequency
   change thresholds for a given system. The thresholds will differ from
   system to system due to differences in processor micro-architecture,
   cache and device configurations.
   In this phase, the user must ensure that no traffic can enter the
   system so that counts can be measured for empty polls at low, medium
   and high frequencies. Each frequency is measured for two seconds.
   Once the training phase is complete, the threshold numbers are
   displayed, and normal mode resumes, and traffic can be allowed into
   the system. These threshold number can be used on the command line
   when starting the application in normal mode to avoid re-training
   every time.


> +
> +* Normal phase. When the training phase is complete, traffic is
> +  started. The run-time poll counts are compared with the
> +  baseline and the decision will be taken to move to MED power
> +  state or HIGH power state. The counters are calculated every
> +  10ms.
> +

Suggest:

* Normal operation. Every 10ms the run-time counters are compared
   to the supplied threshold values, and the decision will be made
   whether to move to a different power state (by adjusting the
   frequency).



> +
> +API Overview for Empty Poll Power Management
> +~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
> +* **State Init**: initialize the power management system.
> +
> +* **State Free**: free the resource hold by power management system.
> +
> +* **Update Empty Poll Counter**: update the empty poll counter.
> +
> +* **Update Valid Poll Counter**: update the valid poll counter.
> +
> +* **Set the Fequence Index**: update the power state/frequency mapping.
> +
> +* **Detect empty poll state change**: empty poll state change detection algorithm then take action.
> +
> +User Cases
> +----------
> +The mechanism can applied to any device which is based on polling. e.g. NIC, FPGA.
> +
>   References
>   ----------
>   

Rgds,
Dave.