XPS 15 9570 Owners Thread

This mess is not new from D ELL. As you can see here... i9-8950Hk in Alienware killed by Dells own low cost models with locked down i7-8750H. Dell have even crippled the Alienware in such way they can't use ThrottleStop same as before.

But FYI. Msi is also a huge firmware and throttling mess. Just ask bro @Falkentyne

 

Yeah but it at least would outperform the XPS for a cheaper price. I borrowed one a few weeks ago and it didn’t throttle nearly as hard. They just sacrifice build quality and screen quality :/

 

I said before several times, that the value in the first row is "bugged"... 2mW drain... lol. you really though that would be!? and if so, how is it then flagged as a drain and how so Windows went into Hibernate because it was a drain??? ... The correct value in the picture is the value 12 but that's also not mWh but 12% overall battery drain over the time of 6h. So the drain was about 2% total battery per hour, or maybe something about 1800mWh. Just because the HW is 18% already shows it is a drain. And that modern standby just doesnt work. How can a software by design hold back the sleep? Hm? That shouldnt be possible, right?

You did a clean Windows install with yours? Maybe I will try this as a last resort when I have time for that. Did you clean the entire disk, like delete all partitions with Windows setup?

My friend bought a new Surface Pro the other day, and guess what? Huge drain after all Windows updates ect. It had even 0% SW 0% HW. I looked on it via TeamViewer and deactivated all services and systrays, though there were just maybe 2-3, like MS OneDrive ect, and the drain was gone. HOW IS THAT POSSIBLE? HM? It should always sleep, regardless of what services or software are open, mh? You cant close everything before you put it to sleep, that makes the whole thing redundant...

 

So I understand you did try to change the power limits and times in Intel XTU? (Don't use XTU and ThrottleStop at once, they fight)
Pressing did adjust these with the previous models; but lately Razer and Apple have issued BIOS updates for this (Razer to remove the i7 cap to 35W, Apple to introduce one because their VRM was't able to feed what was requested assuming the nominal Intel limits and ended up throttling much more). Considering the lack of active cooling it isn't hugely surprising that Dell restricts long-term power. Anybody knows how's this in the i7 9570 (I mean, is also the i7 9570 more restricted than your other laptop)?

Do the throttling time and the power limits shown in XTU match the actual throttling (the package power that XTU or HWinfo64 show and its duration)?

"EDP other" can be a few different things which you can see more specifically in HWinfo64 or XTU. In my 9550 it lights for PL4 when idling, but it doesn't seem to affect performance under load (though I'd be happy if someone could explain what's the problem, this wasn't happening all the time).

 

hi boy, imi help? in particular I wanted to ask if the external video port hdmi, working through the card nvidia 1050, because I want to connect a 4k video of the latest generation ... are undecided between a. external monitor lg 27uk600 directly in hdmi or AOC AGON AG271UG having a razer core x

 

So, just got confirmed from Dell that the second replacement for my previos 2 9570s have been approved. You got to give it to Dell that even though the QC process is abysmal, their Customer Service is really great (at least here in Sweden). Have had some previous experience as well and there was never any fighting needed for what reasonable.

 

I tried using each program alone and neither worked. When using XTU to adjust limits, the power limit change registers in HWinfo but the laptop seems to just ignore my settings. I set limits to normal turbo: 100W, short turbo: 125W, and turbo time to 96 seconds. After two runs of cinebench, I still throttled down to 55w with PL1. I've been looking around at the MSI's throttling characteristics and have found some interesting things. First, the 45w normal turbo power limit usually down-clocks the i7 cpu to 3.6ghz, which for some reason is still higher than my i9's 3.4ghz even though the power usage of these two should be fairly close to each other. Second, these limits are changeable only through BIOS tweaks because Dragon center and some default junk limit the performance artificially. Too me, that means that this is a BIOS/firmware limit that I can't overcome because I don't think you can edit the BIOS of the 9570 to nearly the same degree as the MSI. That would also mean that the i7 version of the 9570 also probably suffers these issues. For all that the XPS has the offer, I'm really tempted to go with an MSi right now because i didn't LM and do a bunch of internal tweaks just to run slower than an i7.

(Just that darn color accuracy and panel quality )

 

They have always really shined in this area compared to their competitors. It makes it difficult to say goodbye to sometimes. I finally had to decide whether the product or the warranty was more important and landed on the product. I would have said goodbye for no other reason that the BGA filth, but there are other reasons as well. They even cripple the performance of their desktops and build them with proprietary crap that complicates things for their customers. Simple service and repairs often requires special proprietary parts that are often not as robust as one might like them to be.

The best warranty is one you never have to use, and the frequency of my need for using the warranty on Dell and Alienware products due to defects and failures eclipsed anything I have ever experienced with any other brand. That being said, a warranty that never delivers sucks pretty bad as well. I've heard plenty of horror stories about MSI and ASUS notebook warranty service failures.

 

> I set limits to normal turbo: 100W, short turbo: 125W

Apple lesson is not to set limits higher than what the VRM can actually produce

> turbo time to 96 seconds

Is the actual turbo time consistent with what was originally written there, and does just extending turbo time have an effect?

> First, the 45w normal turbo power limit usually down-clocks the i7 cpu to 3.6ghz, which for some reason is still higher than my i9's 3.4ghz even though the power usage of these two should be fairly close to each other.

Intriguing!?
Is there a difference in the undervolt between them?

> I don't think you can edit the BIOS of the 9570 to nearly the same degree as the MSI.

I recall folks did some BIOS hacking in the past. But you need someone qualified to do it and find the relevant locations.

> That would also mean that the i7 version of the 9570 also probably suffers these issues.

I wouldn't expect the i9 9570 to be worse than the i7 9570... Can anyone with the 9570 check (LM might not be required)?

> (Just that darn color accuracy and panel quality )

Maybe Dell wouldn't notice if it came back with a different panel

 

@ApplesOfEpicness I find it unlikely that the I9 performs worse than the I7. I find more likely that it will be on par on long term stress testes (on the XPS).
Since you already bought the xps-i9. You have to ask yourself if your typical/real workload is made of short (sub<1min) loads or long heavy multithreaded tasks (encoding).
In the first case, I would keep the i9. In the second case, you might consider returning it and get an i7 based.

Regarding the XPS-i9 vs MSI-i7. Assuming you are interested in long-term heavy multithreaded tasks, I would suggest run prime95 or aida64 for a while on both and see where they settle in terms of temp/Mhz. Realbench is a multi-component test and Cinebench you have to constantly restart it.

 

>Apple lesson is not to set limits higher than what the VRM can actually produce
According to Papusan, Intel recommends these limits for their i9. Regardless, let the vrm deliver as much power as they can muwahahaha.

>Is the actual turbo time consistent with what was originally written there, and does just extending turbo time have an effect?
Pretty much. All I can really be sure of is that it doesn't listen to what I tell it to do.

>Intriguing!? Is there a difference in the undervolt between them?
Nope, they were undervolted the same amount. The GS65 wasn't even undervolted in some situations because i didn't enable throttle stop on start.

>I recall folks did some BIOS hacking in the past. But you need someone qualified to do it and find the relevant locations.
I think I've nailed down a few good placed where i can try editing, but I don't think I can do it anymore since the efi shell they provide doesnt seem to work for me anymore. It also is for the 9560 but hey its worth a shot. http://forum.notebookreview.com/threads/tutorial-xps-15-9550-9550-bios-advanced-options.810307/

>I wouldn't expect the i9 9570 to be worse than the i7 9570... Can anyone with the 9570 check (LM might not be required)?
Well since these seem to be dell set limits in the bios i find the chances for them to be the same highly likely. That said, my i9 is definitely inferior compared the the MSI i7.

>Maybe Dell wouldn't notice if it came back with a different panel
If I could get the XPS 4k screen on the GS65...............

 

@ApplesOfEpicness If you don't have money constraints, why don't you get a razer blade 15? I think they have one with a 4k screen.

 

Does the Dell XPS 9570 still throttle or did someone find a fix?
Still didn't decide wether to go for the XPS 9570 or the Lenovo P52.

Price difference is huge, size difference is there as well.. but maybe there will also be a performance difference.

 

The Razer Blade doesn’t have windows hello support, which is essential for me.

 

LOL! And the XPS has (a working) one with the ****ty camera on the bottom and no extra IR!? ...

 

I’m talking about the fingerprint reader.

 

i9-8950Hk running blazing fast 3.04GHz in Aida64 test (DELL XPS) ain’t pretty to look at... http://forum.notebookreview.com/thr...950h-coffee-lake.810891/page-80#post-10741868

 

Same for me, changing the power limit in Throttlestop doesn't show in HWinfo, but it does with XTU. Either way, doesn't resolve the i9 hitting the per core power limit, and after a couple minutes something resets PL1 to 45Watts (not XTU, it still thinks it's at my 65 Watt value)

Not sure if i'm keeping the laptop (outlet model & has some annoying backlightbleed on the bottom edge), so haven't taped the fan gaps, or re-pasted.
Starting to look like i'll get 3.6GHz stable at ~73 degrees with -.125 voltage offsets (compiling code for 30+ minutes). Temps would be fine, it's just that power limit. Not sure how far i can downvolt this chip yet.

In comparison, the GS73 is struggling to hold 3GHz stable over the same long compile task (thermals, cores hit 93 degrees after 5-10 seconds), might be a terrible paste job on mine but it looks like the their thermal solution is worse than the XPS15's. And the cooling fan (though quieter) is high pitched and makes me want to throw the thing against the nearest wall :-(

Would be very interesting to see how far the XPS i9 could go if we had control over the power limit.

 

This is the reason I got the i9. As a software developer I rarely do anything that requires more that "shorter bursts" buy when I do I want to be as quick as possible. If I needed to do video editing, massive computations etc the real solution would be to get another computer rather than another CPU.

Sent from my SM-N950F using Tapatalk

 

And add to this no TB3 charging, no onsite warranty etc and it is an easy choice for a work machine, but for personal use, a Blade might be better.

Sent from my SM-N950F using Tapatalk

 

I've also considered the P52 but as always, Lenovo was very late to the party. Also, the cooling on that is said to be much less optimal than it should be considering difference in size. And you also have to pay a high premium for Quadro (and anyone that needs Quadro would not consider the XPS15 anyway). Like the idea of a "professional laptop" though and like the Thinkpad design as well.

Sent from my SM-N950F using Tapatalk

 

You can bump up the power a little in XTU but your system will just ignore extreme requests.

Try bumping up say 5W, 10W, 15W then 20W . . .

 

Has anyone considered the 7530 instead of the XPS15?

Sent from my SM-N950F using Tapatalk

 

It even ignores small bumps. The default firmware limits are 56W for PL1 and 72~78W for PL2.

 

Dbrand titanium skin

 

I received my XPS 15 cnx97009 yesterday and noticed the following while setting up.

1. Laptop is currently quite warm when using Google Chrome (bottom feels warm) (see point 3)
2. It seems like that the keyboard lighting on U, H and J seems somewhat dimmed when compared to the Y or O buttons... Is this normal or should I contact Dell to get it checked?



3. I feel like the temps are pretty high while watching a 1080p video in Google Chrome for 1h and 30 min.


Can someone let me know if the mentioned points are normal for a laptop such as Dell XPS?

 

I still see the same behavior, it's all good until *something* else writes to PL1 and then i'll start throttling.

On the bright side, just had a IBT run, 6 cores at 3.9GHz using PL1 64, PL2 85. When i get some time, will try to disable intel's management software in case that's what's touching PL1

 

But so far it seems like the 8850H doesnt really throttle in the P52. And that is together with the P3200 which is comparable to a GTX1060.

But price and size is a big difference between the P52 and the XPS 9570. Which is why I wondered if the XPS is finally throttle-free.

 

- Google is an advertising company. They do a good job with advertising but their free products are generally not very good (albeit they are free). Try another browser.

- That said, Dell (and many other laptop makers) do a poor job applying thermal paste. You will see many people here reapply a better quality thermal paste.

- You can quickly undervolt with ThrottleStop. This takes just a few minutes and requires no disassembly.

- In ThrottleStop you can also enable SpeedShift and choose a high EPP (0= max performance, 255=max energy savings and maybe cooler operations)

- You can also reduce temps by opening "Dell Command I Power Manager" and selecting " cool" in the "thermal management" tab.

- And on some versions of M$ Windoze, try "power options" maybe "balanced" or "power saver".

Finally, this is a very thin laptop so cooling is a bit compromised. Dell certainly could have improved thermal engineering but after the above user tweaks, these perform closer to what Dell marketing is pitching.

I don't see any real difference in the photo. The lighting on these keyboards is not "top-tier" so none are perfect. If the lighting bothers you contact Dell or just return the machine if you are in a return window in your country (30 days in some places). Dell might offer to change the keyboard but note that they will need to really tear apart the laptop as the keyboard is hard to access; and a new keyboard will certainly not be perfect. Your call.​

 

Thanks for taking your time and in-depth information.

I will try to apply some of the points you mentioned, but not all of them. I am kinda worried about the warranty when it comes to opening the backplate and doing work internally on hardware such as re-applying thermal paste and/or changing any other hardware.

I've noticed that the power plan with the default Windows installation that comes with XPS 15 is pretty limited to "Dell" and "Balanced", I will probably get the default power plans back after installing clean Windows 10 on the laptop and use "Power Saver" instead of using "Dell" and/or "Balanced" power plan.

I will report back once I applied some of the points you mentioned.

 

Repaste is pretty easy so you can consider that later if you like.

All the other thermal tips above don't require disassembly and should get you a significant improvement in thermals. The 9570 allows great undervolts so really benefits there.

Go to "create a power plan" on the left side and you can add "power saver"

Also, with the very hot weather around the world this month, your room might be hot, significantly reducing the cooling efficiency of your laptop. Room temps rising from say 20C to 25C have a big thermal impact thin laptops.​

 

Have you found any reviews on the P52 confirming that? Have looked for them myself but so far almost no reviews (like for most workstations). Only found some forum discussions there it is hard to separate facts from opinions...

 

After 32 days I received my 9570 xps 15 with 8750H/16gb/500gb 1080P full HD version.

I actually got the Samsung NVMe ssd :O

But I got some not very good backlight bleed.... In middle bottom and all the edges. Not very nice at all!


Before I fresh install windows. What should I do to test its working properly?


Did update stuff in dell assistant support and fans was going ****in jet air plane for 15 minutes.


Had to wait 32 days. So I just can't deal with another month without laptop. School starts next week! And I got the Samsung ssd and not some hynix crap. So I do want to keep it. Just want to make sure it's working properly before I decide next move

 

Wrong post #fail

 

I know the XPS 15 has a system factory image. What is the proper way to back that?

Sent from my SM-N950F using Tapatalk

 

Tried Dell OS recovery to download a backup of the system image. It seems it only downloads a windows image, lacking drivers/app.

Miss the Dell backup and recovery days.

Sent from my SM-N950F using Tapatalk

 

Speaking of undervolting, I’ve read that the core and cache offsets should be the same and that throttle stop just takes the lower value. However I noticed that my clocks increase if I lower the core voltage but keep the cache the same (my cache is unstable past -120mv). I’m going to see if I can see any voltage changes in hwinfo, but just curious.

Also on a side note, my fingerprint sensor won’t work when I try to log in after waking up from sleep. Anyone else have this issue?

 

A lot of people have noted newer CPUs likely require core & cache voltage to be matched. Some people on these 9570 threads observe they don't need to be matched. There is no recent Intel documentation to confirm so difficult to be 100% certain.

Not sure I follow. Do you mean reduce the undervolt (e.g. increase the voltage)? Or maybe with an increased undervolt (e.g. lower the voltage) the laptop can sustain higher clocks without throttling?

Sleep is buggy.​

 

>Not sure I follow. Do you mean reduce the undervolt (e.g. increase the voltage)? Or maybe with an increased undervolt (e.g. lower the voltage) the laptop can sustain higher clocks without throttling?

Reduce the voltage thus increase the under volt. That reduces the power draw which means I can run higher clocks within the PL throttling.

>Sleep is buggy.

Fantastic... guess I’ll use hibernate.....

 

Hi, my 9570 hit a problem a few days ago in that it was blue screening about 10-15 times a day. Rebooting it would result in a blue screen, 4 times out of 5.

After uninstalling/reinstalling all possible culprit drivers and speaking to tech support, I was advised to do a fresh install of Windows, which I did.

Blue screens have gone (yay) but one thing is not as it was. I cannot access the Nvidia control panel. The option has disappeared from right clicking the desktop. And trying to open it from Geforce Experience does nothing. I have uninstalled and reinstalled the Intel and Nvidia drivers but no change.

It's not the end of the world. I think both GPUs are working and both show up in Device Manager. I'm inclined to accept that this is just how it now is. But does anyone happen to know of a fix?

 

I cannot for the life of me figure out how to update my LiteON drive. Dell has posted 2 diffrent firmware updates and whenever I try to run the update, I get a quick window frame that opens and then closes. I've tried running it in compatibility mode, safe mode, and even trying to update it through hardware manager. Nothing.

Anyone have similar experience?

 

I don't know what the program is and I'm not able to check, but if the message is in a command prompt window try running it from there, that way it won't disappear.

 

Try uninstalling Intel and NVidia video drivers using DDU and reinstalling them again.

 

New firmware for the Hynix SSD available.

Sent from my SM-N950F using Tapatalk

 

Yes, I had similar problems and had to do just that to get NVidia installed again. Still getting frequent crashes, boot problems with my second 9570. Hard to believe that this is like the third or forth release in the same XPS series...seems to be going downhill fast in some regards.

 

Hahaha. Maybe not downhill but a rather predictable pattern.

Dell eventually got even the 9550 largely sorted out. Just took a lot of BIOS and driver releases. And some user tweaking. . .

 

Thinking of ordering the Dell xps 9570 at the end of the month, specs:
i7 6core
16gb ram
1050ti
512gb ssd
FHD

Getting the machine for mainly college work i.e. coding, some adobe programs like animate, photoshop and dreamweaver. Its going to be a huge upgrade from my laptop now and I just want to make sure I'm making the right choice with this one. Any recommendations or advice? Thanks

 

So, I tried updated my LiteON through the USB boot utility and when I did that, the DOS utility told me "no ATA drive found." My suspicion is that the default RAID setting (no Idea why) is causing issues with the readability of the LIteON drive for updating purposes. That's my armchair analysis at least.

And, considering that, from what I've read, RAID -> AHCI requires a reinstall, not really up to doing that.

 

Reading things like this make me wonder if I should return mine and get something else. I've bad no problem so far (except for some reason I can't get the wireless mouse to work) but then I haven't used it much yet either.

 

I have only seen the 4K in person, can someone that has seen both in real life comment about their differences?

The 4K sounds like a dream but since I will use it mostly for work/programming I'm thinking about going with the FHD due to having 3-5 more hours of battery life which is really important to me.

Is FHD sharp enough for a 15 inch laptop?