Increased power usage after sleep

[egore24]

Hello. I am using your latest commit of OC (0.5.8) and I have the same machine specs as you except for I am using the excellent BCM94360NG card for wifi and bluetooth. Previously I was using a DW1560. The issue I am having does not appear to have anything to do with the mentioned cards, as the issue has been present with both.

Issue: After awaking from sleep, tracking power usage with IPG, usage before sleep is 0.73 PKG and after sleep increases and remains at a minimum of 1.9 PKG. Utilization also increases from 0.63 to no less than 1.5. I cannot find a cause for this increased power and resource demand.

It appears to be a power management issue that somehow changes after wake from sleep. Any ideas?

Thanks much.

P.S. This behavior also occurs with Clover.

[sunrez]

@egore24,

I can corroborate a change in PKG power levels after sleep under some conditions (I'm using Clover) : #151 . My PKG power level change is not so significant and in proportion, DRAM power levels drop almost equivalently.

[bozma88]

I have found a repeatable way to debug and find the culprit, but we need @the-darkvoid to interpret this!

(I am on 7500U 8GB 9360 FHD)

Before sleep: 0.6w
After sleep: 1.0w

Enabling-disabling thunderbolt in BIOS makes no difference.
Card reader disabled in BIOS (it adds another 0.5w if left enabled).
CPU base frequency, core and RAM power consumption are the same, even if total package power varies pre-after sleep.

What I just discovered, and is 100% repeatable:

• Boot laptop. Power consumption is LOW
• Sleep laptop. Then wake. Power consumption is HIGH
• [THIS IS THE IMPORTANT PART] Sleep laptop via OS ACPI command* AND IMMEDIATELY BREAK SLEEP before it completes the process by pressing a key / the trackpad
• Laptop will return to woken state but power consumption will be LOW!

* Breaking a sleep-entering cycle is only possible if putting the laptop to sleep through the OS. If you put the laptop to sleep via a low-level command (FN-INS) it's not possible to interrupt the cycle.

What does this mean?
That there is likely a device that, upon wake from sleep, gets restored to a high-consumpton state from the BIOS ACPI tables.
When putting laptop into sleep, some devices are put to sleep by the OS, others by the BIOS.
By interrupting the sleep before it completes (maybe before BIOS ACPI tables have a chance to run), maybe you don't trigger some ACPI code that incorrectly restores power state to a specific device.

Please try this and, if it is repeatable on your systems too, we can continue searching for the cause.

Thank you!