XPS 15 9570 Owners Thread

The dGPU "throttling" happens without Intel DPTF installed too:




But it is never catastrophic. I always thought this whole deabte was about, that it was going into "shutdown" mode aka 800MHz if 75°C was reached. This is not the case. The throttle is minimalistic, to around 1400, more 1500-1600MHz instead of 1700. With an 0.875V undervolt, this was even more stable shifted to 1600-1700 moft of the time, as seen in the graph, over 40 minutes Heaven benchmark. I even added one prime95 thread.


So having DPTF installed or not is kinda redundant. The CPU will throttle anyway because of PROCHOT build inside the CPU.

The shutdown mode aka 800 MHz, actually happens, when the CPU reaches it's TDP limit aka ~45W. This laptop cant do CPU 45W + dGPU 70W at the same time. But CPU 25W + d
GPU 70W it will do. Then just the 75°C throttling will happen, regardless Intel DPFT is installed. Causing it to throttle down a little bit and jump between 1400 and 1700.

There is actually a way better way to just disable turbo for games, which cripples FPS in most modern games. Instead of deactivating turbo, which causes the CPU to use about 10W at 2.2GHz (it is good for low CPU load games though), you can actually cap it at around 3.3GHz (which I found was a great compromise for most games, via TS) and then use around 25W, and this gives in most games I tested again proper 60FPS and results in a way lower vCore around 0.7-0.8 and way less heat, instead of letting it automatically boost and then use 0.9-1.1V >=3.6GHz.

 

Hey @maffle, thanks for this, good to know Intel DPTF does nothing. I am trying the same thing now, capping the wattage of the processor, rather than the GHz, seems to work!

 

I don't have to cap my CPU to play games on my X1E. It keeps the GPU at max clocks all the time in Far Cry 5, unlike the ~1300MHz in the XPS. While the ~450MHz isn't an end all situation, it still needs to be fixed by Dell...

 

@ custom90gt Agree. It is totally stupid to cap it at 74°, 78°C was a good logic value for the XPSs thermals. I dont see any logic argumentation to set it to 74°C. It is just disgusting how ignorant Dell is and stubborn.

@ Mulgul Yes. You can either limit the turbo wattage, reduce IccMax (havent tried so far), or, a bit easier and no risk forgetting the numbers (doesnt work always though), is to do this in the TPL Window:


If you do this, speedshift will actually try to prefer this value. It is not a guranteed "cap" or force, but it will work with 90% of the time. Depending of the load of the CPU. In most games it works great. You can easily play with it around, seting it to 29, 30, 31, 32, 33 and see how it effects FPS/vCore/wattage in game. As I understand it, it is a kinda "try this value" but if the load goes over a certain value (maybe 50%?), it will be ignored, and it boost to full boost again. This doesnt happen with most games though, just where it is needed like short high load spikes, and then this is actually a even better way to limit the thermals, then just a flat wattage cap. This is actually also great for limiting it to for example 8 if you want to watch a movie. 32 also is a great value for browsing which wont trigger the fans but gives way better speeds than turbo off.

 

I have tried a couple of weeks ago to disable DTPM and do some quick tests. I didn't do anything extensive, just disabled the drivers and the associated services and used prime and heaven.

My impression is that the 74c GPU throttle is not controlled by DTPM, it's something imposed at BIOS level.

To be frank, I didn't any difference but I didn't test much.
This is maybe due to the fact that I use Throttlestop's new feature "disable and lock power limit" that might negate some of DTPM effects.
DTPM *might* trigger something based on the Ambient sensors to reduce the "skin temperature" when on battery or on the lap.
After talking with unclewebb, I realized that there are other layers of Power Limit control for the CPU, like PL3 which is activated and locked by Dell (probably at BIOS level) (see @maffle image with 840C8A00).

I don't think the CPU and GPU are connected through BD PROCHOT.

Take whatever I am said with a pinch of salt because I didn't test much.
The laptop works fine for my needs and I don't have time/motivation to dig further, but I follow the thread.

 

@abujafar Why are you using then the "disable and lock power limit" feature of TS? Does it give a little bit of better boost on battery, or what? I am actually just using the laptop 24/7 on AC never on battery.

 

I tried this but I cannot replicate the behavior you described.
If an EPP value is NOT set in the main throttlestop window (checkbox: unchecked), whatever I put in the TPL/Speedshift section is ignored.
If I set a custom EPP value in the main throttlestop window (checkbox: checked), the value that I put in TPL/Speedshift becomes a HARD limit, not a suggested value.

It has nothing to do with the battery, and actually, the right place for that toggle is under the TPL window (it will be moved in a future release).
From my understanding, what it does is: disable (one of the layers of) power limits and prevent something else from re-enabling it.

 

@abujafar For me, it works fine, under the parameters I described. Of course you have to set a EPP value with this to work first. 84 is the only "proper" working really, because Windows will overwrite (seems 84 is the default value of Windows for balanced [on AC?]) it and TS working against it (can get into tiny conflicts if it doesnt overwrite it quick enough, but the time windows are tiny). So I suggest using 84 always. Then for example you set the speedshift max to 31. This then results in a 3100MHz clock (IF the usage is low enough? (for most games that is the case under normal operation)). It is not a 100% hard cap (I noticed several fluctuations with it ignoring the value for a short burts), but an "always working" cap. I am not sure what the rules are, but the max value is broken/ignored sometimes under high load. You can easily notice this for example, if you open Chrome/tab and for a second it jumps to 3.8GHz but then settles to 3.1 (maybe this is the same time window bug actually happening with EPP, and the MAX value is overwritten every x ms/ns and TS overwrites it too every x ns/ms), same if there is a high load (loading a map or whatever) in a game. I like this option of TS way more, than the wattage or power limit options. Because I can easily set a "max" clock and the CPU will automatically use vCore/wattage bound to that clock ratio.

 

Hello Guys,

Will this laptop work with faster RAM than 2666MHz? I know that is the official limit, therefore I ask who have actually tried to plug faster RAM.

Also what is good deal for this laptop?
I’ve ordered mine from Dell Outlet for $940 with I7-8750H, but I think it is little still expensive.

 

This is the best list I found so far. And you'll be pushing your luck > 2666 or with tight timings.

https://www.reddit.com/r/Dell/comments/8r55ys/9570_working_ram_kits/

I personally found using TS to limit boost to 3.2Ghz (+ undervolt OFC) the sweet spot for gaming, have not repasted or VRM padded and seems to survive 10 minutes of Doom (2016) or Colin McRae Rally without throttling. The FPS difference between 3.2 to whatever it would do without limits is minimal, you're mostly GPU bound even at 1080p. I limit to 3.8 on AC, 3.6 on battery and 3.2 when GPU is active, seems to be the sweet spot for me.

It would be nice if afterburner wasn't bugged, I'd then be able to run it all the time.

Yes I also found that quiet mode borks GPU clocks. Thought my unit was broken until I realised there was a correlation!

 

Wow, that is awkward, I've planned to use at least the Kingston 2x8GB 2666MHz CL15 modules. However that is not in the list, but the 2x16GB version. Those Kingston modules are really helping to boost performance.
The another awkward thing is this i7 CPU seeing all these throttle issues, probably the i5-8300H would have been better buy.

 

I already said several times before, how you work around that. You can run it then all the time, like I do too. Just click PAUSE in right click to the graph, done. Then use HWInfo for monitoring into RTSS.

 

Can anyone help me with MaxxAudio?

I tuned sound wit EQ and MaxxBASS. Soung is VERY good.

But when i restart any pleer like a iTunes, AIMP FOOBAR in a 50% of starting pleer the settings of MaxxAudio application turns off.
But thу slider is on the ON position. If i turn it off and then turn on the slider the sound sittings is retunrning.

This sutiation also attend when i start watching youtube with Chrome.

 

CPU-based SpeedShift should be overriding Windows Power Options as a general matter. If you set Windows Power Options to "High Performance" does that give full control to EPP in your observations? Else please post your concerns to the ThrottleStop Guide here for comments from @unclewebb

 

Until my laptop arrives, anybody can help what chipset use this XPS9570 motherboard? I want to install Windows 7 right away therefore I have to prepare, chipset and usb drivers. It would be also helpful to see a screenshoot of the device manager in Windows 10.

 

When i plug in AC adapter, PCH temp rise to 52-55C. Fan turns on at 2500 rpm.
But laptop stays in idle with desktop.
When i plug off AC the fan stopped in 1 minute and temp PCH does down to 40C.
Is it normal situation?

I use startdart charge, not qiuck charge in Dell power manager.

 

The PCH is relatively close to a charging circuit near the battery connector.

 

Can anyone say - is it exist Realtek drivers not from DELL, like Intel HD and nvidia drivers from Intel and Nvdia?

The latest version of audio driver is 6.0.1.8454 in DELL support.
The latest Realtek HDA driver is 6.0.1.8627.
But the Dell sound have a hardware ID
INTELAUDIO...
Not HDAUDIO

And also a driver that is not compatible with ours laptop??

 

@Cooler-master After close to one year, Dell was so generous today to update the Realtek drivers to 8603 https://downloads.dell.com/FOLDER05...ion-Audio-Driver_GGY72_WIN_6.0.1.8603_A04.EXE

 

Intriguing. I wonder what exactly "Enhances audio quality for providing better listening experience" means...

 

See some interpretations here...
https://www.tenforums.com/sound-audio/5993-latest-realtek-hd-audio-driver-version-181.html

 

Interesting. From a very quick listening test, after installing the new driver, it sounds to me like they inserted a compressor or limiter, that catches high transients in a specific part of the spectrum (somewhere in the lower end). My speakers does not rattle anymore, but now there are sitll a distortion going on. However, the distortion seems to be happening digitally now, and not by actual speaker units rattling around because of bass frequencies. Again, sounds like a digital limiting is happening now. Maybe for the better. Will need to test more.

DPC latency is the same, as could be expected.

 

Nope. That is just the bad MaxxAudio software equalizer, you have to deactivate:

 

I actually tweaked my Waves Max audio and think it sounds way better on the terrible internal speakers when compared to turning it off. To each their own I guess.

So my 9570 has a squeaky space bar. After a quick google search, I notice that its pretty common. Has anyone else here experienced that and if so, did you fix it? I have warranty left, but don't really want a tech to come out for that (even if Dell would send one). I saw a video where you can pop off the space key and apply some lubrication but that seems dangerous. Again, just curios what everyone here has done (I apologize in advance for not searching).

 

@ MSGaldenzi My space bar is not spueaky, it is worse, it makes cracking noises and "flips" back and forth if you press on top and bottom. Squakely noises just make my left ctrl, ESC, del, A and tab keys. Premium garbage laptop.

 

https://www.notebookcheck.net/The-D...-than-the-company-first-thought.407681.0.html
No news, actually...

 

Stumbled upon this writeup on battery drain during Modern Standby on the 9570
https://code.luasoftware.com/tutorials/windows/win10-problematic-modern-standby/

It appears that this power-draining insomnia results from an orchestrated attack of Microsoft upon its user base with explicit intention of depriving their computers of proper sleep. Long read:
https://docs.microsoft.com/en-us/windows-hardware/design/device-experiences/modern-standby
A CS-enabled computer is not allowed to support S3; a windows reinstall is supposedly needed for switchover to/from S3 sleep mode; and Microsoft encourages vendors to drop the functionality required for S3 from BIOS altogether.
User rights are restricted to optionally disconnecting wifi during sleep, and possibly inspecting "offenders" with SleepStudy or even sillier and more cumbersome debugging efforts.
But they are not allowed to disable the OS and modern apps from periodically waking up the computer to lesser dormancy levels and processing whatever they deem necessary.
Additionally, there is a set of conditions required to enter the lowest activity state (DRIPS) that may failed to be met.
It is claimed (without justification) that modern standby is more power efficient than S3, but rather openly admitted that the actual power drain is at mercy of app developers.

Somebody should stop the madness.

Btw from the first link it appears that the Hibernate After wake bug known from the 9550 somehow managed to to survive the evolution process to the 9570 ;(
Anyway, Hibernate After appears to be on the retirement path too.

 

@_sem_

1. the author of the blog isnt aware of the c state bug, I knew that blog before and I contacted him asking him to stop spreaking information leading into problems for others to adobt. just disabling CS on the 9570 is not a solution

2. the quote from MS that you have to reinstall Windows switching from CS to S3 and vice versa is a propaganda lie

3. here is the most funny (upsetting) thing, showing, that MS has giving up on fixing modern standby and knows internally, that is is not working, mostly never will: they "secretly" changed the sleep behavior with Windows 10 (>=1703 I think?), so that if more than 5% battery is drained on modern standby, the device enters hibernate automatically ("HIBERNATE BUDGET"). This shows, that MS has no trust anymore in modern standby themself. You cant easily change the 5% hibernate setting too on your own. They called it "adaptive hibernate";

"Defines the battery drain % that the user is allowed in a standby session. Default is 5%."

https://docs.microsoft.com/en-us/windows-hardware/customize/power-settings/adaptive-hibernate

So no, hibernate isnt retirning, and MS "last resort" option. And theyre too stubborn to acknowledge modern standby is a failure, and giving back S3. What a joke.

Funny thing is, modern standby has no function on the 9570, other than waking up 2 seconds faster compared to S3.

 

> 2. the quote from MS that you have to reinstall Windows switching from CS to S3 and vice versa is a propaganda lie

Several things not just that are obviously Microsoft agenda not facts of life, they are obviously pushing for modern and want to kill S3

> 3. here is the most funny (upsetting) thing, showing, that MS has giving up on fixing modern standby and knows internally, that is is not working, mostly never will: they "secretly" changed the sleep behavior with Windows 10 (>=1703 I think?), so that if more than 5% battery is drained on modern standby, the device enters hibernate automatically ("HIBERNATE BUDGET") ...

It is clear that they want laptops "on" all the time or at least much of the time, but it is also obvious that battery capacities are finite, so there is no workaround for them, hehe.
I meant that they want to phase out timed Hibernate After.
The switchover bug is the most troublesome here, drags on for years, though I'm not sure what percentage of laptops is affected by this roughly.

> Funny thing is, modern standby has no function on the 9570, other than waking up 2 seconds faster compared to S3.

I think they mainly want laptops on all the time - email and chat updates, tile updates, skype calls, music playing, and, most of all, permanent tracking of all sorts (what Android has but they don't).
I mean, it is obvious that they don't want to offer the option of constant unconditional DRIPS sleep, which could likely replace S3.

 

No that is a bug in Windows 10 which was fixed some time ago I think. It is basically that the 5% was exceeded and Windows didnt notice it during DRIPS sleep just the moment you wake it up, bam it goes moronic into hibernate. So if you sleep in DRIPS for >=5% then wake up, it goes to hibernate that instant, instead of during the DRIPS state.

Connected standby "MAY" make "SOME" sense on tablets, but not on a damn laptop. You want to move the laptop from point A to B to C and during that time just sleep and use as little battery as possible, and during that time you have no connectivity anyway.

 

Guys i just noticed im getting a 30fps framelimit when running on battery. Even with high performance profile and performance set to highest quality and vsync off on nvidia panel.

Is this a thing for all of yous or some weird bug. I get it that battery drains drastically with gaming but let battery problem be an issue for me why force me to plug into the wall

Sent from my SM-N950F using Tapatalk

 

Most laptops lower performance when running on the battery because they simply aren't made for such high discharge levels. They do that to prevent the marked wear on the battery (as well as the potentially dangerous risks of high discharge).

 

Thats understandable but has anybody noticed and experienced exactly what im saying

Sent from my SM-N950F using Tapatalk

 

I believe the topic has come up a few times here in this thread.

http://forum.notebookreview.com/threads/xps-15-9570-owners-thread.817008/page-123#post-10790262

 

Thanks man. The framelimit contrasts with game emulators (pcsx2 in my case) and deems them unplayable. It took me all morning to figure it out. That article tells you can probably turn off battery boost although not recommended but what harm can it do for a 15-20 min gaming session

Getting to play ps2 games while on the go was a big selling point for me for get a powerful modern ultrabook. Now i dint see the point if you are outside and cant find a power source. Is laptop gaming supposed to happen only indoors

Sent from my SM-N950F using Tapatalk

 

@therock003 Deactivate battery modus in Nvidia Experience, which limits FPS to 30 on battery. I would recommend remove Nvidia Experience anyway if you dont need it and then deactivate Nvidia Telemetry.

 

Sure no problem but what about the quick battery discharge possibly causing an inflated battery that other user mentioned. Could that become a reality

Sent from my SM-N950F using Tapatalk

 

@therock003 No. For your use case I dont see any problems. Just deactivate it.

 

Is anyone use equalizer for Windows 10 in 9570?
I installed Viper and EqualizerPRO. But both did not work in Windows 10.

 

There's no statistics available to us, and I don't recall any report of a battery inflating soon after relaxing battery boost to 60 FPS.
But surely some of these batteries do tend to inflate over time even with regular use, and high-current drain very likely makes it happen sooner. So depends on your priorities. If the lifespan of your battery is the most important to you, avoid gaming and heavy processing on battery altogether. If you need to get heavy work done or you can't live without the game, you'll get a new battery whenever time comes.

 

I'm not sure if what you describe refers to his sentence regarding Hibernate After
> "There is one obscure occurance the notebook wake from sleep but fail to enter hibernation, thus draining all the battery during the night."
It seems exactly the hot bag situation I've debugged with the 9550.

 

Hey forum, a question here.

Since the XPS 15 and Precision 5530 have the same chassis, hardware and I assume fans, how come that there are such big differences in average noise levels in the testing here
34db (A) https://www.notebookcheck.net/Dell-...x-Q-Laptop-Review.355641.0.html#toc-emissions
46db (A) https://www.notebookcheck.net/Dell-...orkstation-Review.324572.0.html#toc-emissions

If it weren't for those noise levels I would get myself the 5530. And from what I read the XPS gets quiet loud which contradicts that testing a little.

Any experience here?

 

rpm?

 

I don't think I even have that installed. Where is it? The only thing I can find in the start menu is MaxxAudioMeters.exe which gives me this message when I try to run it: "failed loading meters". Hmm.

 

@annabel_shanderin UWP app in the MS store

https://www.microsoft.com/store/productId/9NB9SRTL2KPT

 

Thanks.

Off now.

 

Hi owners have 9570,

Is it still a good buy for lasting battery, light gaming and multimedia consumption?

Has all bugs been addressed or fix found?

Should I wait for a refresh or buy an alternative?

 

That's a difficult question to answer.

I think it's a great laptop for battery/light gaming/everyday usage.

Having said that there are still bugs (sleep bug, 75C gpu throttling, etc).

If you tell us your use case maybe that will better let us answer your question.

 

All,

One of our Dell audio engineers has created this confidential survey. Please fill it out. This will help our XPS platform team pin point the issues.

DELL-ChrisM

 

thanks for the help

i travel once every 2 months, so portability and lasting batter for coding, writing and watching movies on 8 hour flight is somewhat important.
I play Destiny 2 casually, as well as Witcher 3. I plan to get an external gpu to play CyberPunk 2077 when it comes out.

frankly i prefer a 14 inch laptop, but MSI ps63 does not have TB3, and Lenovo X1 Carbon is from a less trustworthy Chinese company.

so i am struggling between XPS15, XPS13 (give up gaming), or wait for Asus Zenbook s UX392 (MX 150) to come out.

while i am leaning towards XPS 15, reading about its issues scared me, hence what i am here.