Re: [PATCH v3] soc: qcom: icc-bwmon: Update zone1_thres_count to 3

[Pushpendra Singh]

On 6/16/2026 5:09 PM, Konrad Dybcio wrote:
> On 2/27/26 12:10 PM, Pushpendra Singh wrote:
>> From: Shivnandan Kumar <quic_kshivnan@xxxxxxxxxxx>
>>
>> Reduce zone1_thres_count from 16 to 3 so the driver can lower the bus
>> vote after 3 sample windows instead of waiting for 16. The previous
>> 16‑window delay (~64 ms) is too long at higher FPS workloads,
>> causing delayed decision making and measurable power regression.
>>
>> Empirical tuning showed that lower values (e.g., 2) made bwmon behavior
>> jittery, while higher values (4–6) were stable but less responsive and
>> reduced power savings. A value of 3 provided the best balance: responsive
>> enough for timely power reduction while maintaining stable bwmon
>> operation.
>>
>> Significant power savings were observed across multiple use cases when
>> reducing the threshold from 16 to 3:
>>
>> USECASE			zone1_thres_count=16     zone1_thres_count=3
>> 4K video playback	   236.15 mA                  203.15 mA
>> Sleep			    7mA			        6.9mA
>> Display (idle display)	  71.95mA			67.11mA
>>
>> Signed-off-by: Shivnandan Kumar <quic_kshivnan@xxxxxxxxxxx>
>> Signed-off-by: Pushpendra Singh <pussin@xxxxxxxxxxxxxxxx>
>> ---
> 
> Since this patch does not seem to be moving on its own, I will gladly
> hijack the discussion..
> 
> I was wondering if we could transplant the bwmon devfreq+governor+memlat
> setup from downstream. It seems to be way "smarter", in the sense that
> it keeps some hysteresis information, unlike icc-bwmon, which is purely
> reactive to immediate signals, in addition to (in my brief understanding)
> sampling other data sources, such as the PMU 
> 
Thanks for the suggestion.The downstream solution explores a broader policy space with additional inputs and hysteresis. It is an interesting direction, though evaluating such an approach would likely be a separate discussion.
For this series, we are focusing on a small tuning of the existing icc-bwmon behavior. Based on our latest validation, reducing zone1_thres_count improves responsiveness to decreasing bandwidth demand, allowing DDR and LLCC votes to be reduced sooner and resulting in measurable power savings.

Thanks,
Pushpendra