Nvidia’s GeForce global PR director Ben Berraondo said the issue impacts less than 0.5 per cent of the affected GPUs, with one fewer ROP (Raster Operation Pipeline) than specified.

 He assured that the average graphical performance impact is only around four per cent, with no effect on AI or compute workloads. The issue has since been corrected in production, and impacted customers can contact their board manufacturer for a replacement. 

Affected users will be eligible for a replacement, according to the company. 

While the number of faulty GPUs is relatively small, this is yet another problem added to Nvidia’s growing list of headaches with its latest high-end graphics cards.

The launch has already been plagued by driver issues, including ongoing blackscreen problems that Nvidia is still investigating. There have also been reports of melting power connectors. 

The flaw was not limited to a single manufacturer. Reports of missing ROPs have surfaced from multiple board partners, including Zotac, MSI, Gigabyte, and Manli, as well as an Nvidia Founders Edition card.

 Users who suspect their card might be affected can check using GPU-Z to verify whether it displays the correct 176 ROPs. If fewer are shown, a replacement is likely necessary.