Discord’s RTX GPU Memory Clock Bug Now Has A Fix
Discord’s RTX GPU Memory Clock Bug Now Has A Fix
Likely caused by the recent update which introduced AV1 support on streams, the bug caused some NVIDIA GPUs to drop its memory clockspeeds.
Reddit and Linus Tech Tips forum saw users posting about the bug, and NVIDIA was immediately notified of the issue. The company has since released a statement through its Customer Care Twitter, stating that the bugfix has arrived and users can simply restart the PC for the bugfix to work (the update is done automatically through app profile – meaning no user action is needed).
GeForce users can now download an app profile update for Discord. This resolves a recent issue where some GeForce GPUs memory clocks did not reach full speed w/ Discord running in the background. The update automatically downloads to your PC the next time you log into Windows. pic.twitter.com/89nwugWQFF
— NVIDIA Customer Care (@nvidiacc) February 3, 2023
The latest Discord update introduces a major feature: AV1 streaming support for NVIDIA’s RTX 40 series GPUs. However, the update seemed to cause issues solely on the RTX 30 series models (two reported incidents use RTX 3080 and RTX 3060 Ti), which saw the card’s memory clock drop by 200MHz. RTX 40 series owners has so far no reports of such bugs, however.
The update allows the high-efficiency, open-source AV1 codec streaming for RTX 40 series GPUs, with resolutions up to 4K60 at 8Mbps bitrate (with Discord Nitro subscription). The feature will fallback to standard H.264 format once it detects a user watching the stream running devices without AV1 support.
Source: The Verge | Tom’s Hardware
Pokdepinion: A rather odd bug, but the effects seems to be largely minimal at least.
