XPS 15 9570 Owners Thread

Well, its hard to say with absolute certainty but nothing I've seen indicate that this was improved by upgrading Windows. Disabling Wake on AC and Wake on LAN in BIOS however I think had some improvement. But I have to give it a few more days to make sure what I am seeing is not just something temporary.

 

Thanks. Sigh.

 

Anyone with i9 order that still has not shipped?

Sent from my SM-N960F using Tapatalk

 

Haven't we just had this discussion . Mine is still waiting to be built.

 

Something weird has been happening to me suddenly: whenever I plug my AC into the laptop, my CPU speeds are pretty high at just idle: usually floating north of 2.9ghz constantly. Anyone know what's going on?

 

Haven't we had this discussion also?

My CPU now goes north of 4GHz as soon as there are more than perhaps 10-20% cpu load.

Sent from my SM-N960F using Tapatalk

 

You guys are trigger happy Speedy Gonzales kinda of folks...

You like to go to the max as soon as any activity comes up.

 

I have gotten back an email I wrote last week to a Dell customer support person, mentioning all important current issues. This is what I got back, quote:

"
1) Re-enable S3 sleep mode – nothing further can be added to this. I have stated our stance from Dell engineering and can only reiterate their statement “these systems do not support S3 state”. Whilst I appreciate your points on this, Dell looks at it from an alternative viewpoint and I am unable to change this.
2) Excessive draw on usb/TB when devices are connected in sleep mode – engineering are aware of this and are currently investigating the issue.
3) Audio crackling / acpi issues – again, engineering are aware of this and are currently investigating the issue.
4) Lowering the temp – if engineering have lowered the temp threshold with a BIOS update it’s possibly to protect the GPU. I shall feed this back to them
5) No post error codes – the term processor failure is simply descriptive and to differentiate it from other motherboard related error codes. In this instance it’s more than likely that the motherboard itself has failed. As per my previous mails, I did offer a service call to address the freezing and hanging issues. That offer is still there if you require a service call.
"

This is really disappointing regarding S3 sleep under Windows. I dont want to make any allegations towards the CS person. He mostly just makes his job. If he has instructions by Dell, to tell me this or if he doesnt understand the issue 100% clearly, I cant judge.

The statement, that the 9570 doesnt support S3 is just wrong. If he had the instruction by someone else, to say this, I dont know, or if he doesnt understand the issue. Of course, the 9570 supports S3. They even fixed half of the problem with bios 1.5.0. Linux, for example, uses S3, so it even supports officially S3 under Linux, and would make no difference under Windows of course.

I would welcome, if more people of this thread, who understand the issue, also reach out to Dell, and try to explain to them, that the issue was mostly fixed with 1.5.0, but there still is a bug related to the USB devices.

If we're lucky, and Dell somehow finds out or releases a fix for 2), this maybe will automatically fix the S3 issue too, but that is just a small possibility.

 

@maffle which problems do you currently encounter by manually enabling S3 sleep (using the registry hack) with 1.5.0 firmware?

 

The problem I described before on this thread and on that Intel one too.

< 1.5.0:

- CPU stuck in C2 after S3 sleep under Windows
- package power of about 1.5W under idle (monitored via ThrottleStop)

>= 1.5.0

- C2 stuck after S3 is 100% fixed
- however, the package power is still about 1.5W in idle, even with C8 states

This is mostly caused by a bug, which I can see is related to USB. If you have for example a Logitech USB dongle connected, or use Bluetooth with a mouse or keyboard (which also is USB), and you come out of S3, there is some kind of glitch or stuck driver. You can see that in ThrottleStop on the core0 c0% activity. Where all should be 0.5 - 1% on idle and are before S3 sleep, after S3 wakeup, there is now a 8-12% c0% activity on core0, which mostly causes the 1.5W package power. This is an indicator for a driver (or by the EC) runaway (some infinite loop).

You can see the broken activity on this picture:



This is how it should be:


If you now unstuck all USB devices, this means also deactivate BT, the stuck runaway activity disappears some seconds later, and package power goes back to a proper 0.5W.

The runaway c0% activity wont be seen in Windows Task Manager, just in ThrottleStop. So it is mostly some intern "talking" to the CPU, by lower drivers, or the bios/ec, which Windows wont show in the process list.

 

About the S3 and Dell's reply - the way I understood this reply, reading between the lines, is not that the system doesn't "support" S3, but that they simply don't want to go back to S3, and that's it. It just isn't an option for them.

I actually think it's more likely that Microsoft wants to phase out S3 in Windows (in fact, it's actively doing it) and has told the manufacturers that they'll gradually stop supporting it in new Windows 10 updates, and manufacturers have no option but to follow. XPS 9570 was released with S3 just because that was a part of the "old" BIOS, and, in Dell's eyes, they have "corrected" this "mistake" by removing the "old, redundant option" in the BIOS update.

No new laptop models will come with S3, only modern standby. Microsoft wants to retire the old sleep, and is pushing the new modern standby.
The fact that the modern standby is still not fully ready (and buggy or not fully compatible on many systems) is a whole different topic. Give it another year or two, and it'll be good.
But for now - that's just the way it is, and it's up to us to try to deal with it. Not fair and not right, but that's how it is.

 

@splus Yes, of course, it is that. They don't want to, or MS told them they have to, or the intern structure of Dell doesnt understand the issue (one document says this is a modern standby device, someone other reads it, thinks doesnt support S3, ect). They "designed" (actually they did nothing, the 9560 would have been able to support modern standby too already, the 9550 mostly too) this laptop to be a "modern standby" device as a Windows device. They mostly have some intern "moral/business concerns" or contract with MS, that this is just what they want. The device supports S3 obviously. And as a buyer, you have to conclude from this. There are still other options on the market. Lenovo changed their mind and gave back S3 for some laptops, and it is also working on their machines.

Yet, the XPS are officially coming with Linux too (and Dell gives Linux support), and Linux uses S3. So that argument doesnt hold here.

Also MS cant "force out" S3. They always have to support S3 as a backward compatibility option.

Want to know how awesome modern standby is again? Just yesterday, I put the XPS to sleep with modern standby with AC connected. There was just normal stuff running, some Chrome tabs, some videos in them. And guess what? The laptops fans spun on every 2-3 minutes for 1-2 minutes. Went off. Then again, and again, and again, and again. This happened for about 1 hour with the lid closed. The temps of the laptop were about 44°C when I closed the lid.

 

MS and modern standby. When I spend 2-3000 USD on a system, I bought it, I own it, and I want to control it. If you don't give me that, I will buy elsewhere and recommend others to do the same.

 

So, with all the problems we are having with the XPS15, what realistic alternatives do we have if we should consider something else? I did a try with a Thinkpad P52 and as a computer it kind felt superior in most regards. However, weight, size and design kind of felt like I wanted to call the 90s and tell that I found their missing computer But from a performance, DPC etc situation the P52 is just better.

The only alternatives I can think of are the ThinkPad X1E and the ZBook G5 Studio


https://www.notebookcheck.net/Lenov...GTX-1050-Ti-Max-Q-Laptop-Review.335608.0.html
https://www.laptopmag.com/reviews/laptops/hp-zbook-studio-x360-g5

Where the ZBook is tempting but a bit concerned about weight and size on that that (the none x360 is supposed to be a bit lighter but still a bit heavier than the XPS15 I think).

Or is it so that the 9570, with all its issues, is still the best alternative?

(The Thinkpad P1 seems to be basically the same as the X1E in this regard)

 

Thinkpad X1E looks very good. Extra thunderbolt port, better and spill proof keyboard (and especially the cursor keys, which is, in my eyes, the worst thing about the XPS).

But reviews mention X1E gives only 5-6 hours of battery life, and it's just not cutting it for me. It's not just 80 Wh vs. 97 Wh battery, but XPS somehow has better battery life per Wh. If you don't care about battery life then yeah, X1E seems like a better laptop.
Oh yes, and if you don't care about the price difference!


Unrelated, but I just saw this:
https://www.neowin.net/news/next-wi...ke-your-pc-faster-following-spectre-slowdowns

I think this is a great news. God knows how much of all the weird CPU behaviour and glitches come from the small kernel glitches because of those patches...

 

They are so much more money though then the xps 15!

 

At the end I sold my xps at a reasonable price. Not going to deal with these issues anymore.

 

HP has been a corporate disaster for well over decade, struggling with terrible leadership, strategy, fraud & finances. They really lost their way from the innovative, world-class edge engineering shop they once were. I doubt HP are really focused on producing world-class laptops.

Dell is what it is. Relatively cheap laptops with half-hearted cowboy engineering. I really like my 9550 but Dell is not willing or able to fix all the quirks.

Lenovo has produced some excellent laptops; if you have critical secret data, I'm not sure I would trust a Chinese brand. Albeit, any laptop you purchase is likely sourced from China lol.

The big non-BGA laptops with open-BIOS are really the performance kings. Downsides include size, design, and price. But they seem to deliver what they promise so are a good value in my view.

 

what about asus?

 

If I remember well for you the GPU performance was more important than CPU?
If you want a thin and light laptop with a good GPU, then why not go for Aero or Razer? They should be both pretty good. It all depends what you want.

As for brands - you can always have a bad luck with any brand. Or good luck and no problems. But in general I don't think Asus has a good reputation in regard to quality and support, at least not in the last couple of years. Razer was real bad, but it looks like they're turning it around.

 

I tried out the Razer Blade 15 as well and while it sure had its merits, it just did not work out as a work computer. Proper keyboard backlight, finger print scanner, fiddly scanner etc. But there was quite a lot to like as well.

 

I agree aboit the Razer if you primarily need a work computer. From his previous posts I understood the GPU was more important for him, so in such a scenario a thin and light gaming laptop might be a better option.

For me personally, all things considered, postives and negatives, I really don't see a better option than a current XPS 15.

As a newcomer to the XPS (and Dell) I can understand that many people who had 9550 and 9560 would now almost hate on it, having accumulated some annoyances with the past models, and seeing some things haven't been resolved in the current model either.
But coming from some other models, it seems to me, that some of these issues are really not as big as they are partrayed to be.

In some laptops from other brands these negatives are almost taken for granted and people almost don't even think much about it. They just don't complain about it.
I've never heard anyone complaining about thermal throttling in other laptops, it's not a problem, of course laptops thermal throttle, especially thin and light ones. But if XPS throttles then it's the end of the world.
And reading these opinions before I bought my XPS I thought wow, it sounds like the performance on XPS is really miserable, it throttles far more than any other laptop in history of laptops (based on the amount of negative writing from both tech media and people who know more about it). But then, I was REALLY surprised to see that XPS 15 was at the top in all the benchmarks, short burst and sustained (not counting thick gaming laptops, of course).
Something didn't add up here. Someone wasn't really realistic and objective.
Especially seeing how people give a pass to many other laptops that have the same or worse issues, which are just ignored, or somehow become minor issues.

And then I gradually understood the psychology of it - grass is always greener on the other side. People start picking on smaller faults and make them bigger, just for the sake of picking on it.
It's like the XPS became tech media's favourite bashing laptop.

I do understand the history and the reasoning for this (especially when coming from 9550/9560 owners). But still, I don't think it's objective and realistic, especially when compared to other laptops, where often even bigger faults are just taken as "Ah ok, it's just the way this laptop is. Well ok then, it's not a big problem anyway..."

 

@splus My main problem with the XPS15 is the poor quality of BIOS, firmware and drivers (at least the "special ones" like docking station etc). That it would throttle I knew before buying it and that I personally have really no problem with.

 

lol. The 9550, 9560 and 9570 are half-baked laptops that with a bit more development effort could have been superstars in the light-and-thin BGA world.

 

Tell me which other 1800€+ laptop has that poor thermal paste job.

Inviato dal mio SM-G950F utilizzando Tapatalk

 

Probably all of them. Dell uses a heatsink with pre-applied thermal paste. I think the easier question would be show me a laptop without thermal paste applied in this manner.

I've opened hundreds of laptops and nothing has been what I would consider good or great. Sadly it's the industry standard.

 

I agree. BIOS is definitely buggy and looks a bit like the BIOS developers don't do any tests, or aren't fully aware that one change can affect other things.
I was trying to portray the full picture, taking all the positives and negatives together, and weigh them against the other laptops. And I still think, ALL things considered, this laptop is better than others in this category.

Yes, for sure it could have been better (and maybe it will get better than it is now, if they iron out some problems). But I think that's more or less true for all laptops.
Still, what I was saying is that, for what it IS, with all its negative points, is still a very good, if not the best laptop, WHEN COMPARED to other laptops in this category (and even in any category). Relative context, not absolute.

I don't know, is that the reason it's at the top in performance charts? And better than some even more expensive laptops?
Don't get me wrong, I'm not saying it's perfect! I'm just saying the grass isn't greener on the other side. And I'm saying that some of the same issues exist in other laptops, yet I don't see anyone complaining about them on those other laptops. But XPS gets all the heat...

 

Dell has a "weekend sale" on Dell XPS 15 that ends today. I'm considering a 9570 to replace my 9550. I very rarely use my laptops on battery power -- raw horsepower for using Adobe Lightroom is more important for me.

The 9570 i9/32 GB/1 TB is $2999, the i7/32 GB/1 TB is $2899. That's only $100 difference.

Is there any reason NOT to spend the $100 for the i9 version?

 

Different question. Any reason not to buy the i9? Answer: Not just any reason, a ton.

 

Apparently i9 throttles more than i7 and has even worse performance than i7, especially for sustained performance. But for apps that need short burst performance might be OK, better than i7.

Maybe there's an i9 owner here or someone who tried it, who could give you an accurate answer?

 

Spend it

Sent from my H8166 using Tapatalk

 

With enough cash to burn, fell free to take the i9, but the performance seems to depend more on your luck on the paste job lottery than on the difference between i7 and i9.
It is questionable if the 9570 is a good direction if you're not happy with the horsepower of the 9550.

 

I would also think that the resell value of the i9 vs i7 would be worth the $100 in the future.

 

That's the main reason I bought the i9. Well, and the fact that I could and don't do "second best" willingly

But I agree that the i9 probably does not make all that much sense from a price perspective....well any perspective to be honest.

Sent from my SM-N960F using Tapatalk

 

Hopefully some good news:

Dell XPS 15 9570's DPC latency issue should be fixed soon

https://mobile.twitter.com/AzorFrank/status/1054193282723573760

 

But, in your experience, for a quick and short burst performance, or basically responsiveness and speed in short actions, wouldn't i9 be much better than i7?

All the reviews always do thorough and long benchmarks (where i9 tends to overheat), but is there any benchmark that gives you a good feeling of performance for short actions?

Photoshop (and actually 99% of apps that need only short actions) should do much better with i9 than i7. But rendering a long video in Premiere probably wouldn't, or would even be slower than i7. At least I guess..

 

That is the logical assumption, or at least that it should be "better" at least. But since I don't have an i7 with similar specs to compare with its hard to say. If you factor in price in the equation also it might tip in favor of the i7 also, depending on what premium one would pay for the i9.

 

Thats great news, although I still consider mr Azor to be a modern version of Bagdad Bob so I believe it when I see it

 

That is no news at all. All tweets by AzorFrank are pure propaganda as proven several times in the past.

 

Like what?

 

You cant help to wonder what computer Mr Azor actually uses himself since it took someone on twitter for him to realize that the XPS has a DPC problem to begin with. Guess that part did not come as a complete surprise to use customers...

 

He's Alienware guy first, XPS second. But all these execs are usually disconnected from practical reality and probably get informed by the guys below them who feed them different story.

Still, he's the XPS boss and his word means something, especially if it's a public statement like this one. Now it'll be a personal embarrassment for him, not just for Dell, if they don't fix the DPC.

I just wish they'd fix other issues as well, especially the high power drain in modern standby and change the low GPU temp limit back to where it was.

 

Well this isn't really news. I recall that Dell legions were on it at least a month ago already. Hopefully they do come up with something

 

I recall he promised improved cooling on the previos version of AW laptops as well, nothing happened IRL.

 

In fact, most of the reviews do relatively short tests where the CPU doesn't get to heat up properly; they also don't tend to load up the CPU so heavily because it waits for other stuff. Because of this, the i9 usually does score better in reviews.
I guess review units are often cherry-picked or manually repasted. If results are unusually bad reviewers tend to try to fix obvious issues firstly. On the other hand, a less sever paste issue may "pass" in some cases, with a big impact on the score. If a review doesn't report results before/after undervolt+repaste, it may not be worth much. I mean, sometimes it appears that the i7 is faster than the i9, but more likely it is within the margin of the paste variation and testing error.

Some reviews like Notebookcheck also do a stress test, with sth like Prime95, maybe combined with Heaven or Furmark, and a tougher game, where one can see what goes on under load. Cinebench is another quite relevant load test, but it doesn't appear to put a heavy continuous load on the CPU. What really matters is the demanding programs you use yourself - for instance if you do photo stuff with LR you likely care for import speed, for editing responsiveness in Develop and for batch export speed, which mostly isn't covered in reviews, you can only guess from the benchmarks.

Mind the i9 is capable to run long-term at higher power than the i7 according to the Intel specs, but this is locked down by Dell.

 

I asked Frank Azor once whether the XPS 15 9570's RAM would be soldered to the mainboard or not like the 9575. He didn't know and had to ask the engineering team.

 

Wrong. They really don't care.

 

Never mind this post the problem fixed itself... lol

 

Well, the number of crashes with the TB16 has gone down significantly the last week, which I believe is probably related to me turning of the Wake on LAN and Wake on AC. Not really sure why though. So, given that and the fact that I didn't really find an obvious replacement for the XPS15, I have cancelled the return of my last XPS15. So as it seems I will still be one of you

 

This is from before the 9570 launch, asked if it will have coil whine.




The "invested a small fortune" is mostly one of the most pathetic things I ever read. Maybe they should better invest in a few cents per laptop of better coils, instead of propaganda lies.

There was another tweet before I cant find right now, where he was asked about throttling on the 9570 and TB performance issues, and he said, they "redesigned the mainboard from the GROUND up" and there wont be any throttling or TB issues. That was also a lie, the mainboard is basically identical with the 9560 (even the 9550) with some tiny little minor changes.

Do you use the LAN port of the TB16? Not sure if this is related, but I know for sure, that the LAN chip in the TB16 has a hardware bug leading to corrupted data and maybe also crashes, which mostly just occurs under Linux of my knowledge. I thought it was fixed under Windows, but maybe it isnt 100%. You can read about it in detail here:

https://bugs.launchpad.net/dell-sputnik/+bug/1729674