We have followed problems with the new GPUs of the NVIDI RTX 50 series for a while, but they now seem to hit the critical mass. It seems many owners of the new RTX 5080 And RTX 5090 The cards undergo accidents and in particular black screens, with some undoubtedly less supported reports of “brick” cards. We interviewed Nvidia about the problems and we were informed that he “currently investigated the problems reported with the RTX 50 series”.
As always with an emerging problem, the current situation seems complex. Most often, owners of the RTX 50 series report black screen problems. The scenarios in which this occurs varies. Some say that this happens when changing the resolution or rates refresh. Others find the black screen kicks under a heavy load, while others associate it with multi-monitors configurations.
Some users then find that the problem persists on the hard restart, the card not detected in the peripheral manager or the system bios, while for others, it seems more intermittent with a restoration of restoring for a certain time before that the black screen does not strike again. As we said, it’s complicated.
Now we have to predict all of this saying that we have had very few problems with our RTX 5080 and 5090 advice so far. We have experienced some accidents when you are in Alt in and out of games to access various reference and capture tools, but nothing at all persistent that remained after restarting. Essentially, we have no direct experience of these black screen problems reported even during the overclocking of RTX 5080 to an inch of its life in silicon.
However, we were tedious about the complete cleaning of the pilot before installation. Hold this thought is important. Anyway, when we write these words, the number of users reporting problems on platforms, including the Nvidia forum (An example of many here) and Reddit (Ditto, here) went well beyond a few isolated and anecdotal incidents.
Likewise, many YouTube videos signaling or having directly encountered the problems are emerging now, including one Jayztwocents which offers an overview and another of Boosted media where they both experienced black screen problems and seemed to have corrected them.
Indeed, we interviewed Nvidia and we said: “We are investigating the problems reported with the RTX 50 series.” So we will have to wait and see concerning official solutions or pilot hotfixs.
Anyway, to cut a long story, the problem most likely seems associated with the latest version of Nvidia driver 572.16. There are “permanent” failures of cards. But at the very least, these seem much more isolated and can well constitute a typical failure rate.
In other words, when you empty a batch of graphics cards on the market and buyers enthusiastically report their experience, you necessarily have some, either by manufacturing problems or user errors.
It should also be noted that some GPU owners of the RTX 40 series also report problems with driver 572.16. Thus, these problems may not be entirely specific to new RTX Blackwell cards.
In addition, in particular with regard to the RTX 5080, we do not think that the problems relate to the material stability or the GPUs executed near the wire to the factory parameters. We have managed a stable overclock GPU of +525 MHz and an overclock of memory of +1,000 MHz with our RTX 5080 Founders edition, which means that it works on a solid 3.1 GHz. It is a very heavy overclock according to recent standards.
Picture 1 of 3
If we are confident enough at this stage, there is a real problem, the next step is a solution. We suspect that Nvidia will abandon a blog post in the near future describing the problem and promising a correction in a short time or accompanied by an update of the hot pilot.
Until then, the best advice is emerging to solve the problem is a full wiping of the driver with DDU (display of the non-stansor driver), then reinstalling the driver. This suggests that the problems arise from a kind of conflict with the driver’s residual data. To come back to our previous comment, it may not be a coincidence that we have directed DDU before installing our revision cards in the RTX 50 series as a routine precaution and that we have not encountered the problems black screen.
The definition of your PEG-16 graphics port on PCIe Gen 4 mode can also be a wise short-term precaution following PCIE signaling problems With the new RTX 50 GPUs, especially if the DDU wiping has not fully succeeded.
Anyway, if nothing else, all that has rather added any weight to the idea that a complete pilot erases before any new GPU installation is a very Good idea.