Battery life improvement?

[Thinkpad4by3]

This is mainly for the X210 but if you have one of the other machines, feel free to answer. How much better is the battery life on the newer machines. I mean the batteries for these machines are getting old and weak, but also the use of new LED displays, more efficient panels, and much lower power electronics compared to what it came with. And if it is better(which I assume it is, how much better is it if you know what you originally got on one of those machines.)

[vladisslav2011]

I have a third batch x210 with 2880x1920 BOE NV126A1M-N52 LCD, NVME SSD and ar5b22 wifi/bt combo.
And I have two batteries:
One bad 9-cell chinese fake, 68Wh. Charges in ~3h, discharges in ~10-12h. It shows average rate:
4.5-5w offline, minimum backlight, light load
5.5-6.5w online , minimum backlight, light load (thunderbird, firefox with some web1.0 pages opened, some PDFs)
The other one is good 9-cell Kingsener battery declared as 94Wh, but it is 98Wh really. It charges in ~4h and discharges in about 16-18h. And it shows other average rates:
5-6w offline, minimum backlight, light load
5.5-7w online , minimum backlight, light load

The biggest power sucker here is an LCD. I'ts backlight draws about 1-1.2w at minimum and about 5.5w at maximum. And minimum brightness is about 20%. The panel itself draws about 1.6w constantly even being in dpms off state. I've tried to run the laptop with physically disconnected LCD (from motherboard connectors) and external display. It draws only 1.6w with external display off, idle and offline.
P. S. I have pathched i915 driver to force-enable PSR even if it's timings do not meet the requirement. It work very well, saves about 0.8-1w power when screen is not refreshing and does not produce any artifacts...

[harryK]

P. S. I have pathched i915 driver to force-enable PSR even if it's timings do not meet the requirement. It work very well, saves about 0.8-1w power when screen is not refreshing and does not produce any artifacts...

Hi,
Is your patch available somewhere? From one of the screenshots that you posted in another thread it seems that enabling PSR also allows the processor package to get to a deeper C-state. Your powertop screenshot shows PC9, whereas I have never been able to get lower than PC8...

[vladisslav2011]

It's here:
--- a/drivers/gpu/drm/i915/intel_psr.c 2019-03-04 02:21:29.000000000 +0300
+++ b/drivers/gpu/drm/i915/intel_psr.c 2019-04-30 21:32:44.830631955 +0300
@@ -589,7 +589,7 @@
adjusted_mode->crtc_vtotal - adjusted_mode->crtc_vdisplay - 1) {
DRM_DEBUG_KMS("PSR condition failed: PSR setup time (%d us) too long\n",
psr_setup_time);
- return;
+ // return;
}

crtc_state->has_psr = true;

It allows the package to reach pc9 even with display on,but I've never seen it reaching pc10.

[L29Ah]

Ok, i think you meant

Code: Select all

diff --git a/drivers/gpu/drm/i915/intel_psr.c b/drivers/gpu/drm/i915/intel_psr.c
index 84a0fb981561..53c811241979 100644
--- a/drivers/gpu/drm/i915/intel_psr.c
+++ b/drivers/gpu/drm/i915/intel_psr.c
@@ -625,9 +625,8 @@ void intel_psr_compute_config(struct intel_dp *intel_dp,
 
 	if (intel_usecs_to_scanlines(adjusted_mode, psr_setup_time) >
 	    adjusted_mode->crtc_vtotal - adjusted_mode->crtc_vdisplay - 1) {
-		DRM_DEBUG_KMS("PSR condition failed: PSR setup time (%d us) too long\n",
+		DRM_DEBUG_KMS("PSR condition failed: PSR setup time (%d us) too long, proceeding anyway\n",
 			      psr_setup_time);
-		return;
 	}
 
 	crtc_state->has_psr = true;

[harryK]

It's here:

Code: Select all

--- a/drivers/gpu/drm/i915/intel_psr.c	2019-03-04 02:21:29.000000000 +0300
+++ b/drivers/gpu/drm/i915/intel_psr.c	2019-04-30 21:32:44.830631955 +0300
@@ -589,7 +589,7 @@
 	    adjusted_mode->crtc_vtotal - adjusted_mode->crtc_vdisplay - 1) {
 		DRM_DEBUG_KMS("PSR condition failed: PSR setup time (%d us) too long\n",
 			      psr_setup_time);
-		return;
+	//	return;
 	}
 
 	crtc_state->has_psr = true;

It allows the package to reach pc9 even with display on,but I've never seen it reaching pc10.

Your patch doesn't work for me, but I have a different screen. The 1920x1200 screen probably doesn't support PSR at all...
Another PM related issue: could you enable L1 substates for nvme drives? I tried with a couple of drives, but if I enable it in the bios the system locks up upon restart

[leedatsing]

I have a third batch x210 with 2880x1920 BOE NV126A1M-N52 LCD, NVME SSD and ar5b22 wifi/bt combo.
And I have two batteries:
One bad 9-cell chinese fake, 68Wh. Charges in ~3h, discharges in ~10-12h. It shows average rate:
4.5-5w offline, minimum backlight, light load
5.5-6.5w online , minimum backlight, light load (thunderbird, firefox with some web1.0 pages opened, some PDFs)
The other one is good 9-cell Kingsener battery declared as 94Wh, but it is 98Wh really. It charges in ~4h and discharges in about 16-18h. And it shows other average rates:
5-6w offline, minimum backlight, light load
5.5-7w online , minimum backlight, light load

The biggest power sucker here is an LCD. I'ts backlight draws about 1-1.2w at minimum and about 5.5w at maximum. And minimum brightness is about 20%. The panel itself draws about 1.6w constantly even being in dpms off state. I've tried to run the laptop with physically disconnected LCD (from motherboard connectors) and external display. It draws only 1.6w with external display off, idle and offline.
P. S. I have pathched i915 driver to force-enable PSR even if it's timings do not meet the requirement. It work very well, saves about 0.8-1w power when screen is not refreshing and does not produce any artifacts...

10-12h is very impressive for a bad 9-cell 68wh battery. My brand new 6th gen X1C with a 4K screen can barely do 7 hours.