[Guide] Improving Battery Life on Windows [+Enabling Deeper C States]

I checked again and removing the hub(USB mouse and a cooling pad are connected to it) did allow the package power to go down to 0.6W instead of 0.7W. USB Selective Suspend doesn't seem to be working.
Regarding PCI-E Suspend, do you mean PCI-E ASPM? That is disabled and I could not find much info regarding it.

 

Exactly that. There's a Lenovo thread discussing it.

That CPU package state is pretty close to the lowest it'll go. But PCI-Express ASPM will drop non-CPU power.

I know Intel platforms can afford to do better because ARM Windows systems do it. Also, their Tablet oriented Z-Atom platforms did comparably to ARM devices.

 

@IntelUser Is this http://forum.notebookreview.com/thr...ws-for-saving-2w-of-power-consumption.606100/ the one you were referring? I checked with HWInfo and ASPM doesn't seem to be enabled only for the Intel iGPU. Considering how little power it consumes when idle I wonder if its worth the effort to get it working? Getting USB Selective Suspend to work properly might be more beneficial I think.
Edit: After a clean boot min. package power fell to 0.5W which as of now is the lowest ever. Still can't get USB Selective Suspend to actually work.

 

Do I need to get my cstates sorted? Definitely not majority in 7 8 9 when not doing anything intensive at all.
brand new spectre x360 with intel i7-1065G7/


https://imgur.com/Es0n9le
I've undervolted it a bit as the fan very easily turns on and the thing gets really hot so hoping optimise it a bit better to keep it cooler and quieter

 

I think you are fine, as it has C8 active. Just see if you can lower the impact of background applications somehow so its in the lower states more often. Lower the C state the harder it'll be able to reach. What's your package power?

Honestly for Icelake, idle power is excellent. The problem is that the process/uarch is still immature and the voltage/frequency scaling is poor. Hence when its anything above idle the battery life suffers. The leaks for Tigerlake looks much better but we'll see.

Yea that's what I'm referring to. The iGPU is controlled by internal mechanisms and can be power gated so don't worry about that. The iGPU doesn't use the PCI Express to connect to the rest of the CPU either - its all connected using the ring bus.

 

1.2 to 1.5 W when idle.

I just sat and watched it for a few minutes when idle and it didn't enter C8 at all.

 

@Griffin3567 Your image is showing C8 at 16% of the time. It looks like it can go lower, but is it at least going into C7?

Your package power should be able to reach under 0.5W. Beyond that it'll have a minor impact, but 1.2-1.5W is quite high.

 

Che0063 thank you for the guide, previously managed to achieve 0.5w pkg power on my 9750H gigabyte aero 15, but had to reinstall everything due to an accidental wipe. I've been trying for a day, installing factory image and also a fresh copy of the latest windows, updating my drivers regardless it's Stuck in C3. Reading your C-states guide again, I uninstalled the Intel RST drivers and BINGO!

 

Hi, I just registered here to show my thanks for this guide, and to the results here.
I recently bought an Acer Aspire 7 A715-73G with an Intel i7-8705G CPU. I read decent reviews about its battery life. However, it could barely do 4 hours in idle (or well, which I thought to be idle then). With a 48 Wh battery, that means 12 W in idle, I think we can all agree that's absolutely unacceptable.

I was this close to RMAing the notebook, but I found this guide. I have to note that before trying everything here, I also tested the battery life in an Ubuntu live USB, and even that showed tremendous improvement. Previously in Windows when streaming 1080p videos it used ~18 W, and in the Ubuntu live install that decreased to ~9 W. This is without any powertop autotuning or tlp installed.

Now when I rebooted into Windows to start the guide, for some unknown reason the power consumption was much much lower than before. I could get 2.2-2.3 W in idle on 0% brightness, 2.8-2.9 W with 30% brightness, which is absolutely amazing. With that, I could finally reproduce what various reviews showed me. After that, I did most things in the guide here.
I undervolted the CPU by -150 mV, maybe I could go lower, but that's enough for me. Package C states are mostly (80%+) in C8, which is pretty cool.

Now after a restart, some update got installed with Windows 10 and in idle it's 4.8 W, slightly more when browsing the web. I find that a bit weird (why did it increase?) but I'm satisfied, with this the laptop is decently portable. I still wonder why this happened after booting into Ubuntu. Maybe some drivers got updated just before that? I have no explanation really.

I attached my ThrottleStop settings here if anyone with a similar laptop or CPU has anyone problems.
(EDIT: Just saw that the C State windows covers it, but I also have "Disable Turbo" checked in TS. Also the temperatures are normally around 40 °C, I just ran a benchmark before the screenshot)

 

@andriii25 Wooot! Congratulations! Though I can't think of a reason why it would do that? Maybe Linux smacked Windows on the head and forced enabling all power saving features?

I kid!

Unrelated news, but Intel announced Core processor with Hybrid technology "Lakefield". It should reach new levels of low power and battery life for x86 devices.

Modern standby should be far more effective as well with its C10 power state being rated at a mere 2.5mW. That's a 12x difference from Amberlake Y's 30mW figure.

I'm interested in getting one if they make a device in the proper form factor. 2.2lbs 40WHr+ battery, convertible please!

 

i did all this but cannot reach under 4w with windows7.
i get 2.5w with win10, any ideas what might cause this?
im running a i5-6300u only integrated HD520, al drivers up to date.
throttlestop is saying 1.5w pkg power on win7 while 0.5 on win10.. no ideas anymore.

 

Windows 10 has better support for the latest power management features. Actually, it was 8/8.1 that did this and 10 is an offshoot of that so.

 

i have another laptop with an even newer 7200u and is running under 3w in idle without problems on win7..
maybe another idea whats missing, some update maybe for ahci or something similar?

 

Even if idle can be lower, Windows 10 should in general achieve better realistic battery life. Windows 8+ introduces improved timer coalescing to allow devices to be in idle for greater periods of time, even under "load".
https://arstechnica.com/information...-on-the-inside-under-the-hood-of-windows-8/2/

 

In the never ending quest to reduce power consumption, I'm now working on my 2019 Spectre x360. i can get extremely low idle power consumption (0.3w), but it seems to spike up higher than expected when actively doing something on the laptop. Just scrolling through reddit, it really peaks power consumption to between 8 watts and 13 watts. I have the cpu undervolted by 0.115 mv, and everything is set to more power-efficient settings (ie cores are allowed to park, hardware duty cycling enabled, speedstep enabled, turbo disabled, etc).

Does anyone have some suggestions for where to look to investigate a non-squared frequency to power ratio? I mean going from 800 mhz to 1.3 ghz on the cores pushes the laptop from consuming between 2.6 and 3.5 watts up towards 8-13 watts. Anything from typing to scrolling causes this dramatic increase in power.

EDIT: I mean total power consumption, not just cpu power consumption. I get an increase from 0.3 watts on the cpu cores to around 2 watts, but this doesn't account for the rest of the change in power consumption of the laptop. I don't see markedly more activity on the igpu, and I can't measure the uncore power consumption because it spikes it when I record it.

 

@Fighwithadeagle Maybe you shouldn't disable Turbo on your device?

Icelake's base frequency is very low so with Turbo disabled it'll only be at 1.3GHz. If non-CPU power dominates as you say, you'll benefit from Turbo since it'll finish tasks faster.

 

Base frequency on. 9750h is 2.6ghz, but it's not even coming close to 2.6 with automatic frequency adjustments. I even have speed step to set at 220 to favor battery and set a matching frequency in windows. This is all on a power saver energy profile

 

I see, you are talking about the larger Spectre, not the smaller one with 10th Gen 10nm.

Can't really say. Maybe its not really running on the iGPU and it engages the MX250?

At least it has a huge battery.

 

Is anyone having issues with the new windows update, version 2004? After updating I am no longer able to access anything deeper than c3, battery life has tanked by 50%, power consumption is up to 15w from about 7w.

 

Hmmm not really; I've been using 2004 for a few months now. Try rolling back and see if its a driver change/incompatibility, which it most likely is.

 

I tried rolling back a few things but couldn't figure out what the issue was, it ended up being the intel display driver. My system uses optimus and something must have broken when windows updated me to the new "intel command center" version of their driver. It wasn't broken until I updated windows so I didn't suspect that one. I'm running the manufacturer supplied intel driver from 2017 now.

Edit: I was close on the gpu driver, I thought it was the intel driver, but after a couple restarts nvidia was updated to the newest driver and the problem presented itself again.

So for anyone with an optimus notebook, does Nvidia driver 451.58 break c-states for you? I had to roll back to 446.14 to get them to work.

 

I currently have a i7 8750h laptop....whats the lowest I can get the power usage in mw so far the best I did and without effort was 9500mw drain...it has a 65wh battery and would be nice to get 10 hours...anyone?

 

I think you're dreaming a bit expecting 10 hours. I bet you can easily achieve 4 hours, maybe up to 7 with browsing, etc. My laptop is a bit older than yours but has an i7 7700 in it, same size battery, and I average 5 hours of productivity usage.

The only notebook of this spec I know of that gets around 10 hours has a 91wh battery.

https://www.eluktronics.com/MAG-15u

http://forum.notebookreview.com/thr...element-fusion-15-mag-15-vapor-15-pro.830272/

I would expect the Mag-15u to do slightly better than the normal Mag-15, but owners are saying they get 8-9 hours with that massive battery.

 

Got a Razer Blade Stealth 13" (2020) recently and reinstalled LTSC on it right as I got it and use my live script to configure it for maximum efficiency. Took quite a bit of work getting the laptop into C8 (And eventually C10).

Now I'm sitting at 0.3~0.4w idle with 1065G7. Basically It came down to:

- Literally enable every (chipset) function / device possible in BIOS (Basically totally reverse as to optimizing a desktop PC where you'd try and minimize complexity/complication by disabling everything you don't need). Kind of logical I guess as most devices start in their high power state and only drivers will put them into low power state after OS booted)

- Install all and the best drivers, since Razer doesn't deliver ANY drivers for this new notebook, I had to eventually figure out everything myself and grab the latest from wherever I could find it.

(This excludes Intel Dynaming Tuning as it really is more of a acpi / chipset negotiator and only saw problems from installing it like high acpi.sys dpc latency, and gave zero benefit both on idle or stress in various workloads). So I was left with 5 'unknown device' entries all related to DTPF.

- Once all drivers were optimized on top of bare LTSC, I finally achieved 0.3~0.4w idle with 40+% C10. After that it was still troublesome to keep it there after having to install (mandatory due to TDP unlock when on AC power) Razer Synapse 3 software which basically installs some drivers to be able to communicate with the keyboard lighting / other system functionalty. So yes third party vendor software can easily screw things up here. (Also the software forces a 1ms timer vs 15.6ms so thats another bad thing.)

Basically had to make a script to auto unload/load all Synapse stuff in the background based on AC/DC condition.

- Unhidden all possible powercfg options and tinkered with those, a few 100mw extra saved with AHCI link power management to 'Lowest' (Funny as i have no AHCI devices attached, still seems to help), enable core parking and some other parking adjustments so it only uses my first 2 cores unless really necessary. NVMe timeouts obviously but dont really affect package C states in my experience.

- Last but not least ThrottleStop and undervolt Vcore, Cache, SA and iGPU to -80 -80 -60 -95 (mv).

Results are that the laptop basically remains cold as if it were to be off, I mostly get 18~ hour of battery indicator just idle'ing (Keyboard / screen brightness on like 20%

Some other observations:
After S3 standby the C-states don't want to recover and get stuck, only solution I found is to never use S3 but purely use hiberation (suspend to disk), this is not really a nuisance as it only takes a few seconds to resume vs a second in real standby (altho this will depend on your bios firmware).

This looks to be related to Intel Thunderbolt, firmware issue or a combination of both. (I've seen similar reports on Dell laptops also with thunderbolt issue causing this)

 

Another case of this. I really think this issue affects more people than we realize. I do not have a thunderbolt port and I experience the same issue. If you ever find a solution to this please update this board, I'd love to be able to use sleep on my notebook instead of hibernate.

 

Maybe you have thunderbolt (in chipset) but simply no physical port or connection to it.

A few times I had luck and C8/C10 started working again somewhat after disabling/enabling thunderbolt device in device manager a few times. But this was more sporadic and couldn't get any consistency in it. Also didn't go as deep.

Also tried a handful of thunderbolt drivers from various websites.

 

You need to get this info out intho the community - check if there is a reddit Razer thread and make a PSA there - You, nobody, should have to deal with a broken core functionality on a premium notebook. If this doesn't get solved within a few months by a BIOS update, return the notebook.

Also, thanks so much for your detailed report. It's nice to see what goes on under other peopeles' laptop.s Also, you seem to be one of few users who actually have C10 Package states enabled. I wonder if you would mind checking if your laptop panel supports Panel Self Refresh? AFAIK that's the thing that needs to be enabled for C9/C10.

I get what you mean about the manufacturer software decresaing c state residency. Lenovo Vantage (the equivalent of your Synapse) also has multiple background processes to monitor and change stuff like backlighting etc. If you're interested, you may want to see if you can directly control your hardware settings via e.g. the Embedded Controller with RW-Everything

Use Synapse to change settings, and see if any thing here changes with it. You might be able to directly edit settings from here (and thus create scripts)

 

Yeah I actually didn't find much info or people talking about this on the Razer reddit itself, maybe it simply isnt the right community for this kind of talk. Also doubt many people there care about much other than gaming (while these ultrabooks are nicely suited for any type of work and, apparently also a very nice battery life for this powerful beast.

Also the iGPU/Panel indeed support Panel self refresh, and I guess Razer did support decent enough firmware to even allow C10 propery (even if their own software handiwork seems to neuter any deep package states, kind of hilarious )

I noticed that when my windows configuration script processed my installation and after reboot my Cstate residency was worse, turned out to be one my .reg files that showed seconds on the taskbar clock, which together with Throttlestop updating was enough to bump my milliwatts up by a few hundred!

Also checked with RW-Everything a week or two back, and I could not find any sensible changes except for the 3 power profiles which affect CPU TDP duration.

Balanced, Battery and Gaming flip an address value to 1, 2 and 3 when I change them over.

However so far no luck when I manually change this value with Synapse not running, but havent spent too much time on it yet.

Maybe its a lingering kernel driver that prevented me from altering anything effectively or I might be missing something here.

I really dont care for Synapse bloatware, even with a minimal forced offline installation and hacking off some junk its still memory/cpu hungry.

Fan control on auto is fine and default, and the only other option is 'fixed' (really who wants to manually change their fan speeds all day long?) and i dont care that much for kbd backlight customization either. The default when no driver intervenes is slow rainbow cycling which is totally not distracting.

I do wonder how they change the fan speed if not through embedded controller tho, unless I just havent found the value yet. Scripting that would be amazing as auto is kinda conservative on hot summer days.

I wonder if theres some better debugging software to see what kernel drivers send to where, i bet the razer kernel drivers just basically gets used to process commands from synapse.

 

Do you have any way to point me in the right direction to find this out? I don't know enough about thunderbolt and now I'm wondering if the gpu implementation could be close enough to a thunderbolt port that it causes the same issues. Again, this is totally over my head at this point, so I could be way off track.

 

Very interesting thread I was hinted at just some moments ago on Reddit. Why didnt I know about this thread before... when I have the energy and will power in me, I thought make some longer post on me, with my experiences with all kinds of "c state drama", modern standby and S3 sleep on my Dell XPS 15 9570, I found out and went through over now close to 2 years. I think I maybe have some hints to give too to some other people. I encounters countless firmware bugs on the XPS laptop, some might be related, some not. With the loss of all package c states, with broken c states after S3 wake up, with not working C10 state, and so on. Also made the same experience on this laptop, some people are reporting on here, of the broken C states after S3 sleep, and that it seems to be related to TB or USB controller. I made a old post about it some time ago here:

https://www.reddit.com/r/Dell/comments/91313h/xps_15_9570_c_state_bug_after_s3_sleep_and_modern/

I also recently found this old Reddit post of people posting of their issue with c state related to the PCIe controller and the dGPU:

https://www.reddit.com/r/Dell/comments/civtn0/g5_5590_rtx2060_kills_cpu_cstates_after/

I can so far at least describe 3 or 5 different "c state anomalies" on the XPS 15 9570.

I noticed about the broken c states after S3 wake up, that if you remove/disable all USB devices, important that includes the Bluetooth controller too, that the c state bug goes away after 2-3 minutes. This didnt seem to work 100% the time though for me.

This is my current state of the laptop on battery:


As you can see it works in that current state pretty good.

About C10, at least for this laptop, it is never reached if you have power connected, c10 would just work on battery. Sometimes it also just randomly stops working, and after a Windows 10 reboot, it works again.

On the XPS 15 9570, there was some nasty bug introduced with bios 1.14 which totally killed off all package c states:

https://www.reddit.com/r/Dell/comments/eak4hh/warning_do_not_update_to_bios_114_on_the_xps_15/

I found out by pure luck, that if you load bios factory settings, the package c states went back working again. Totally have no idea, why. There was a similar firmware bug in the past with the 9570, where suddenly VT-d and VT-x support stopped working after a bios update. And again loading factory settings made it work again.

For the "c state bug" after S3 sleep, I kinda found out, that it is not really a c state bug, but more of a "runaway activity" of around 12% on one core, after waking up from S3 sleep. You can see it on this screenshot:



Process explorer never hinted to any process running, always showed 99% idle, so it must be some loop/runaway in the firmware itself, or in a driver or maybe even the TB controller/firmware causing this. Disabling all USB devices or removing them, sometimes caused this to go away and back to normal 0.5W but not every time. Important is, that Bluetooth device also is USB and has to deactivated for this to untrigger.

Despite of working C10, and a 100% DRIPS state, this laptop consumes around 1100mW/h of power during modern standby or around 1.1%/h, which is just horrible.

According to disabled ASPM, which is also disabled on all Dell laptops and tablets, I thought this was because of it wont be compatible with NVMe SSDs and could lead to data loss in combination with modern standby. I think it doesnt matter if it is disabled on those devices, if the SSD properly enters its own lowest sleep state, and also all PCIe devices do.

For people who want S3 back on Windows 10 2004, this might be a hint btw: https://www.reddit.com/r/Dell/comments/h0r56s/getting_back_s3_sleep_and_disabling_modern/

 

I don't know about that, my laptop idles prettty quickly with less than <0.25% cpu cycles after like 30 secs to a minute after resuming from Standby. I don't have that kind of runaway usage here or whatever its causing it.

I can do whatever, but unless I start screwing around with thunderbolt device in device manager I'll never ever recover back into C10, until I actually hibernate again and resume.

If the dGPU actually causes issues that probably indicates that PCIe ASPM is not working properly. It should work fine in general, also for NVMe but they also turn themselves down.

No clue about modern standby, my laptop doesn't support that but yeah 1.1% an hour is horrible, I can hibernate my razer and comes back out 100% after like 18 hours of not having it touched hehe. Unless you need it, better disable it imho.

 

@Emtee_ and with what did you measure the "<0.25% cpu cycles" you are speaking about? Like I said above, the issue is not shown in task manager or for example process explorer. Both would show 99% idle after S3 wake up. Only TS would hint at this runaway issue, showing a c0% activity on one random core, most core0, which other task tools wont show. I dont know why this is the case, but I assume this hints that this is a process running in the firmware/EC or in a driver (mostly the usb driver or TB), which wont show in process managers. ASPM is deactivated on most laptops and tablets, so that is kinda redundant. I dont accept using hibernate, that is a total no-go and just nonsense for a laptop. Hibernate wake up takes about 15 seconds or so on the 9570 with 16GB of RAM.

 

I see, my hibernate literally takes 2~3 seconds (Ok, not much running the background atm but still acceptable)

ASPM disabled? Mhh.. Isn't that basically PCIe Power Saving in windows power plan setting? As that has a definite effect here in terms of heat production.

I just looked now and S3'ed again, all my cores are instantly below <1.0 C0% except core0 which remains a little above 1.0 sometimes.

 

That sounds not right and more like S3 wake time. 2-3 seconds hibernate wake time? I highly doubt that. Hibernate wake up would mean a cold boot, so it is not just the entire boot cycle, it is also load of the RAM from the SSD. That cant take 2-3 seconds. S3 wake up is already 2-3 seconds.

So this c0% activity after S3 sleep may just be a specific bug on Dell laptops, or just the 9570. I tested it on 3 other 9570 and they all showed the same issue. The 12% c0% activity then obviously would lower the c states and result in around 1.5W drain. Like I said, deactivating all USB devices cleared this broken state sometimes. So it may be the same issue, but manifesting differently on yours.

No it is not. There are 4 power states of PCIe also link speed.

 

Well my C10 states tell me the contrary after resume from hibernation, so does my laptops led indicator which stays lit up (breathing) in regular S3 standby, whilst hibernated the unit is fully off, and still have 99 to 100% bat remaining after more almost a full day of not having it used, versus S3 standby where I can clearly see a few % have been depleated.

Also S3 resume is merely a second if not shorter, what else is it taking 3~ seconds to resume then? S3 doesn't show the circular loading icon either which only occurs on a cold boot or resume from hibernation.

Last but not least, it takes a tad longer to go to sleep as you can clearly see it has a delay writing down the contents to the disk.

If a cold boot barely takes 6~ seconds (measured right after post to login screen appearing), why is it so hard to believe that suspend from disk only takes a few seconds tops? Its quite possible these days.

Remember this is a brand new gaming grade notebook, and the POST on this laptop is incredibly fast as well, next to my pretty optimized windows 10 install.

Edit: Ok I remeasured from literally pushing the button to login screen, its roughly 4~ seconds including full POST ;-)
It it acceptable for me? Yea sort of, is it acceptable in general? No. Someone did a sloppy job no doubt 1 second versus 4~5 seconds is still a huge difference in retrospect. Battery wise I doubt it makes much of a difference, but it does depend on the interval on non-usage.

The situation on my desktop is this: 4~5 seconds post, for various (stability) reasons and otherwise I dont want to reduce this (although I can), and resume from hibernation right after POST is even faster than my new notebook, looking at 2~3 seconds tops, but this system is also way faster.

Pretty sure last time one of my external PCIe cards did not play well with PCIe ASPM, I fiddled around with the bios options (Changing ASPM to Native, or bios controlled) and in which case when it was on native I could disable ASPM through the OS in my power plan through this very option. Not sure, this is just my observation.

It is often more of a nuisance and totally not worth it on a desktop but on a laptop I can not really understand why it would not support ASPM, even my Intel NUC does and that is theoretically just a very small form factor PC.

 

Because waking up from hibernate is cold boot / POST PLUS loading the hibernate file to ram. So it cant be shorter than the 6 seconds, 6 plus 6 seconds sound about right, resulting in 12-15 seconds. I just did a hibernate wake test with 7GB of RAM usage on the 9570. It took 23 seconds to wake up from hibernate to login screen. After 6 seconds the pre POST was done, after 12 seconds Windows was beginning to boot, took then another 10 seconds to load the hibternate file. Mostly another reason I love Dell laptops, not.

 

Bleh, that's long. Sorry to hear I guess hibernation is no solution for you due to insanely long total loading time.

Doesn't look like that laptop is that cheap either.

Just curious what the sustained speed on your disk is? My laptop and desktop both around 3.6GB/s sustained (luckily hibernate file is barely to not fragmented if pre-allocated 100%)

On regular usage I rarely go beyond 10GB mem usage anyhow.

 

@Emtee_ I use a Samsung 970 Evo in that laptop.

 

Emtee, if your laptop wakes up nearly instantly after opening lid, then it supports modern standby. Ever since 4th Gen U core, Intel chips support what's called a S0iX state, which is a combination that tries to bring the benefits of S3 like low power and S0 like wake up time. The support is not perfect though. I think the reason is Intel chips still have a way to go in lowering S0iX power.

Supporting C10 is a key part of that, but as @maffle device indicates the U series uses a lot less power in all states than H series chips do. Also, the U chips have an on-package PCH while H chips need a separate one.

 

I just randomly now again had this package "c state bug" with my XPS 9570 yesterday, and I have no idea, whats triggering it.



All was working normally for days, lots of hibernate wake ups in between, and *poof* yesterday just that. I noticed it actually because of fans always running, looked into TS, and saw the permanent 2.5W usage and all package c states off... even with 0.4% c0% activity.

No USB device on the laptop, no TB. Had to reboot or shutdown to make it work again.

Before reboot, I tried to disable and enable again several devices in device manager, like the Nvidia GPU, wifi, bluetooth, with no effect.

I am using the latest Intel SATA AHCI RST driver right now, version 17.9.1, may it be a try worth using the default MS SATA AHCI driver? The Intel AHCI driver isnt really necessary at all, right? Any disadvantage not using it?

 

What Nvidia driver are you running? 446.14 works for me, 451.67 nukes c-states on my notebook. Windows now has some sort of GPU scheduling built in with the 2004 update. I suspect the new Nvidia driver tries to do something with that and it doesn't play nice. If you're running 451.67 revert and see if it fixes your problem. If that isn't it, then maybe do some investigating about windows 2004 and gpu stuff.

https://www.zdnet.com/article/windo...first-nvidia-now-amd-supports-gpu-scheduling/

 

I used to be under the impression that the Intel ACHI driver always was better than the MS Sata driver - but even then, I noticed on my Xiaomi that even with the MS driver, C states were working fine. I'm not sure. Yes, Windows 2004 introduces WDDM 2.7, which might fiddle with GPU scheduling.

 

Im on 451.67. The c states work normally for a long time, they just randomly get broken every once in a while, maybe once every ... I dont know, every 10 days or so? And then it is broken until I reboot the laptop. The best is of course, that modern standby then obviously also doesnt work when this happens and causes a catastrophic drain.

 

Okay, did reverting to the older Nvidia driver fix it?

 

I dont see any reason to believe it has anything to do with the Nvidia driver.

 

Hi! I’m new to PCs. Coming over from Mac. I bought and have been tweaking my PC for about 3 weeks. I have a MSI GS66 10SFS-032 Core i9 10980HK. I followed this guide besides the “disabling services” part. I’m having an issue with the C State. I checked my drivers and everything seems to be up to date, although unsure because I’m new to all of this. After unplugging everything and using battery mode I’m able to get around 10 hours, but c state seems to idle about 50-60% in C7 with the rest in C2. I turned on PKG Demotion and underdemotion and that seemed to move the idle to C6 about 65-72% and the rest in C2. My CPU usage is at 1% as well. C8-10 seem to be completely inactive. If anyone could help me out please LMK

 

Nope it doesnt support modern standby (Also force disabled it through registry for good measure).

Again the fast recovery from hibernate is simply due to quick POST, combined with fast hardware and neutered LTSC installation. The fact that the BIOS POST is so fast is just pure luck, the fact that resume from disk is so fast is due to the quick Samsung NVMe and DDR4 memory in this beast.

Sadly Razer refuses to fix anything and I get the usual crap from them 'we can send it to one of our repair centers', after they asked me to make a thorough video to demonstrate the C10/standby issue (Thanks for all the useless hard work razer!)

So C10 remains bugged after S3 (Suspend to ram), and I will keep using pure hibernation S4 (Suspend to disk).

 

@Che0063

thank you bro for the outstanding work. With your instructions I have finally been able to reduce my i9-9980HK from 5 watts to ~ 1.4 watts. This now runs in battery mode when working in the office at around 45 °. Before it was over 55 ° and the fans were running.
Now the Dell Inspiron 7590 is perfect. Thanks all!

https://www.directupload.net/file/d/5906/iuaq8zzw_png.htm
https://www.directupload.net/file/d/5906/hk2bwhnt_png.htm
https://www.directupload.net/file/d/5906/egp7bfre_png.htm
https://www.directupload.net/file/d/5906/efmfc5t6_png.htm
https://www.directupload.net/file/d/5906/qa5pzcfr_png.htm
https://www.directupload.net/file/d/5906/3gq9462c_png.htm
https://www.directupload.net/file/d/5906/jpxdbxmx_png.htm

lg.

 

Hi, checking back in with a new Laptop.

I now have the Lenovo Yoga Slim 7 with a Ryzen 5 4500u. Idle package power is at 0.15-0.25 Watts (goes to 1.8 Watts when moving mouse) according to HWinfo. I'm now trying the same old Intel tricks to increase the battery life even further. This thing already lasts a comfortable 12 hours on a 64 wH battery and I can even game with locked framerates for almost 10 hours. This chip is quite impressive but I think there's nothing I can do to improve efficiency further.

 

Small update may help others: C10 actually does work all the time after standby, it looks to be a cstate reporting bug.

Power draw and heat output both confirm this.

(Just in case anyone else is pulling his hair out on this one, check actual consumption/heat)