NVIDIA Investigating Reported GeForce RTX 5090 & 5080 Black Screen & Stability Issues

[T0@st]

Unlucky owners of problematic GeForce RTX 5090 and RTX 5080 graphics cards have posted feedback across various online community sites. Press outlets started to take notice of these documented issues soon after the launch of NVIDIA's debut wave of "Blackwell" GPUs. PC Gamer has "kept track" of reports relating to black screens and miscellaneous failures—the site published an investigative article late last week, following user feedback "hitting critical mass" across Reddit and Team Green's own forum. A request for comment was sent over to NVIDIA HQ; PC Gamer received a brief response. A company spokesperson confirmed that their team is: "investigating the reported issues with the RTX 50-series."

Several PC hardware community members have documented their troubleshooting experiences—the most common suggestion involves downshifting from a PCIe 5.0 connection to 4.0, on the PEG-16 graphics port. Unfortunately, this step did not resolve black screen issues for certain owners—a member of the buildapc subreddit explored a wide array of troubleshooting channels. They re-installed Windows 11 (23H2), adjusted BIOS settings, experimented with monitor connections, and played around with drivers. Best results were produced by connecting a single monitor to their MSI GeForce RTX 5090 GAMING TRIO OC's DisplayPort, with nothing else hooked up to the other ports (DP and HDMI). They suspect that Team Green's GPU drivers could be the source of frustrations; corroborated by a recent VideoCardz news piece. In addition, the 572.16 driver is reportedly affecting "certain GeForce RTX 40-series." A smaller number of owners have discussed a "bricking" of cards; VideoCardz believes that China-exclusive GeForce RTX 5090D models are suffering the most. Manli will be analyzing a "bricked" unit at its service center, in the near future. Colorful did not reply with a comment on the situation.




Skip to the 10-minute 30-second mark to watch the segment covering "GeForce RTX 50-series GPUs Encountering Stability Issues."

Gamers Nexus has asked its audience to voluntarily send in problematic cards, for investigative purposes.

View at TechPowerUp Main Site | Source

 

[GamreCaleb]

Now whos having driver issues

 

[catulitechup]

der8auer find very interesting issues

Nvidia RTX50 with MFG - Multi Flame Generation

 

[Darmok N Jalad]

It's unfortunate that we never do get the stats on these issues. Much like the Raptor Lake stuff, it's just anecdotes, a handful of personal experiences, and the company just pointing at the warranty. It's unrealistic to expect every product to be without issue, and only the people with issues are going to complain.

 

[Solaris17]

You know whats strange, is for the first time in a bit I had a HELL of a time installing 572.16 on my 4090. I know it is more than likely just coincidence but my machine really didnt like them for some reason.

 

[R0H1T]

Gamers Nexus has asked its audience to voluntarily send in problematic cards, for investigative purposes.

You mean clickbait? When do we start calling this sh!t out

GN was among the first ones on the "user error" bandwagon with that 4090 BS & is still riding on that pony!

 

[sbacc]

Apparently you should get de ASUS Astral 5090, because it has per pin sensing, which the FE does not.

So the ASUS card CAN balance the load over all six 12V cables.

Actually, Buildzoid made a video today about this as well, and the Astral can't balance the load, it can only let the user know if there is a power balance problem...

 

[redzo]

Apparently you should get de ASUS Astral 5090, because it has per pin sensing, which the FE does not.

So the ASUS card CAN balance the load over all six 12V cables.

From the below buildzoid's video, it looks more like a detection mechanism for the Astral. There is no way to mitigate it. User intervention is definitely not the way to do it.

This is starting to look more like a recall, at least for the 5090 FE. People will probably stop buying 5090's unless this thing gets settled.

 

[HTC]

Apparently you should get de ASUS Astral 5090, because it has per pin sensing, which the FE does not.

Apparently, nVidia should have made the 5090 with TWO connectors instead of just one: that would have PERMANENTLY eliminated the problem (on the 5090s).

 

[redeye]

der8auer find very interesting issues

Nvidia RTX50 with MFG - Multi Flame Generation

or Nvidia RTX50 with MFG - Multi F***ed Generation

 

[sethsu]

Premium prices with premium problems

 

[Daven]

Ugh, such a long history of Nvidia driver problems. That combined with blind loyalty have allowed such underlying problems to go on for far too long.

ATI had the same poor driver mentality long ago but no more under AMD. It’s time to start holding Nvidia accountable for these problems.

 

[soupinator]

Apparently you should get de ASUS Astral 5090, because it has per pin sensing, which the FE does not.

So the ASUS card CAN balance the load over all six 12V cables.

It actually can't balance the load, it can only detect an unbalanced load because all the 12V lines are still combined into one after the sensing circuit.

 

[Dragokar]

Apparently, nVidia should have made the 5090 with TWO connectors instead of just one: that would have PERMANENTLY eliminated the problem (on the 5090s).

Or simply don't wire all 12V to one shunt.....

 

[HTC]

Or simply don't wire all 12V to one shunt.....

That would only MASK the problem: NOT eliminate it.

 

[TheDeeGee]

It actually can't balance the load, it can only detect an unbalanced load because all the 12V lines are still combined into one after the sensing circuit.

Yeah, seems so. It can only detect a bad connection on each of the six 12V cables. It's something i guess, but not the solution.

From what i've gathered so far, either GPU makers need to stop cheaping out, or PSU makers need to rework the 12V-2x6 circuitry. Atleast one of the devices needs to know what's happening with each of the six 12V cables.

 

[Darmok N Jalad]

I wonder if this is one of the main reasons why AMD didn’t even bother with a premium-tier chip. They probably knew they were going up against a 600W beast, and it just wasn’t worth the risk and engineering effort to compete. 300W is already a lot to deal with, but at least the engineering solutions are well tested.

 

[Dragokar]

That would only MASK the problem: NOT eliminate it.

I know and I dont like that connector design since they introduced it......but since people buying it we still need to deal with it.

 

[dyonoctis]

Ugh, such a long history of Nvidia driver problems. That combined with blind loyalty have allowed such underlying problems to go on for far too long.

ATI had the same poor driver mentality long ago but no more under AMD. It’s time to start holding Nvidia accountable for these problems.

That seems more like a massive blunder, rather than business as usual though...most of the Nvidia drama of the past years were about misleading specs/questionnable business move/sabotage of games. I've rarely heard of news worthy driver issues.

As I said before, you can't build brand loyalty if the customers can't enjoy your product. Even apple was forced to correct their blunders (Mac Pro 2013, butterfly keyboard, removing mag safe, removing ports on the MacBooks pro, GPGPU being straight up dysfunctional on MacOS X...). Loyal customers can forgive occasional blunders, but there's not a single company in the world that managed to survive on a culture of constant mediocrity with products that are straight up not working.

 

[sephiroth117]

You mean clickbait? When do we start calling this sh!t out

GN was among the first ones on the "user error" bandwagon with that 4090 BS & is still riding on that pony!

User error yes but they clearly and extensively stated that the tight margins around 12vhpwr were to blame too

also, this was with 450W. With 575W 5090 new problems, different potentially from what we saw, may appear.

We’ll see how it evolves and how Nvidia responds should the issue becomes bigger