Post reply

Warning: this topic has not been posted in for at least 120 days.
Unless you're sure you want to reply, please consider starting a new topic.
Message icon:


shortcuts: hit alt+s to submit/post or alt+p to preview

Topic Summary

Posted by: Trade
« on: January 24, 2019, 14:21:21 »

Need to update after the new bios 1.7 fixes as you posted news "Dell resolves the XPS 15 9570 NVIDIA GPU issue with BIOS 1.7"
Posted by: Pauwel
« on: December 21, 2018, 14:42:32 »

I have the 9570 and had this bug as well, updated to bios 1.6, no change.
Factory reset bios (not bios reset option) the laptop and the problem is gone for now...

hope this helps someone!
Posted by: JK
« on: November 20, 2018, 14:04:39 »

So whats the update with this?
Posted by: Toka
« on: November 13, 2018, 20:40:33 »

A follow up on my previous post above.

It seems I was wrong - Dell replaced the motherboard today and my 5530 is now working properly. This might be due to the fact that the new board has Bios version 1.2.2.
Posted by: Toka
« on: November 10, 2018, 19:26:53 »

I can also confirm the same issue on the Dell Precision 5530 with P1000 GPU. At first I thought it was a random issue which made it incredibly frustrating and I spent hours trying to work out what was going wrong. Was it the power settings, was it the drivers, was there a setting in the bios that needed to change?

Using GPU-z I eventually found it seemed to be related to the temperature of the GPU. As in the article at temperatures somewhere below 48-50°C the GPU performs poorly, with low FPS in benchmarks such as Cinebench 15, Unigine Heaven 2 and 4 and SPECviewperf 13. You might have thought that when running such benchmarks that in time the temperature would rise enough that full performance would be achieved but unfortunately that doesn't always happen - sometimes it just idles along at a steady 44-46°C doing poor FPS. This seems to be the case for Unigine Heaven 2 and 4 and SPECviewperf 13.
According to GPU-z, below the 48-50°C threshold the VDDC voltage seems to be capped, which is perhaps related to the problem. Above the threshold the voltage capping stops and the GPU operates as expected.

As stated in the article you can make the GPU to perform properly by 'forcing' the temperature up by running another GPU intensive task. You can also use a CPU intensive task to acheive the same result, presumably because the chips are close on the motherboard and share the same cooling system.

I've told Dell but they just want to switch the motherboard which presumably will not fix the problem.
Posted by: Folfy
« on: November 07, 2018, 23:12:07 »

I can actually confirm the observations, just got my XPS 15 9570 from the factory today. Initially I had BIOS v1.3.x (0 or 1, not sure anymore), and furmark went just fine, starting from 42°C or so. Then I updated to v1.5, and instead of running slightly above 1000MHz on the core, it was limited to about 500MHz (power limiting, acc. to the overlay). Once it hit 48°C, magic happens, and the GPU clocks just normal again.

Now in hindsight, I think I should have made some proving screenshots  :-[
Posted by: mkdr
« on: November 07, 2018, 08:35:21 »

I have the exact same issue, though already reported months ago here:

It has nothing to do with bios 1.4.1 and 1.5.0 and mostly preexisted way before that. I was always wondering, why in some games it didnt boost up and kept at 99% GPU load, good low temps, but bad FPS. Then randomly it started to work and then randomly fall back. I never got on the point, that it might be TOO LOW temperature related because they were always good around 45-48°c...

If this is true, it is ridiculous. So if dGPU temps fall below 48°C... it throttles DOWN... regardless of the load it has? This cant be true. But it would explain a lot.

Because I try a lot to optimize my XPS 15... undervolt, set all to low in games, fps cap at 60, and I dont game intensive games anyway. That could explain, that I optimized the laptop "too good", and then it throttles because of this? This is pathetic.

Oh and btw... reduce the damn reCaptcha behavior I had to solve 30 RECAPTCHES BEFORE I COULD POST. Or switch to the new reCaptcha 3.0
Posted by: A B
« on: November 06, 2018, 22:29:09 »

Got my new XPS 13 9370. Since day 1 I started to notice a sign of overheating. It doesn't happen very often but it is a concern for me. Today, it happened again after I set the power mode onto Best Performance. The laptop is plugged in and the keyboard started to heat up with fan running trying to cool down. I think the ventilation is somewhere on the bottom in the back which makes so sense. Thanks for sharing this information..have shared this link with others keep posting such information.
Posted by: Ale
« on: November 06, 2018, 15:42:55 »

@Vladimir - IF they fix this issue. Nvidia has this thing where they pay $ to the manufacturer for every faulty GPU. This covers the vendor's warranty cost but completely screws the consumer.

They also promised us to fix the overheating issue that caused laptops to fail, in the end they didn't, warranty ran out, it turned into a class action lawsuit against Nvidia, and despite Nvidia "losing", the one who ended up losing was the consumer as they bribed the lawyers to turn against the people they represent. So everyone got crappy netbooks to replace 1k+ laptops.
Posted by: reven1010
« on: November 06, 2018, 12:45:50 »

Same thing goes for G5/G7 and 7577 laptops. Same bios, same issues.
Posted by: Vladimir
« on: November 05, 2018, 22:44:21 »

Dell is aware of this issue and they are working on a solution.
Posted by: Redaktion
« on: November 05, 2018, 17:38:11 »

It appears that Dell's most recent BIOS releases for the XPS 15 9570 have caused a GPU bug in some devices. Originally reported on Reddit, some XPS 15 9570 laptops power limit throttle their NVIDIA GPUs until they hit a temperature threshold, causing noticeably worse performance in games.

C 2018 » Impressum     Sprachen: Deutsch | English | Español | Français | Italiano | Nederlands | Polski | Português | Русский | Türkçe | Svenska