Random driver timeouts occur during gameplay on 7900xtx.
Random driver timeouts occur during gameplay on 7900xtx.
aye, I just crashed with the latest Adrenalin version, 30 mins in The First Descendant. Everything at stock. Adrenalin showed GPU HOTSPOT temp was at ~71 oc. Not even 80. Then I went to tune the frequency down to 2704, I think stock was 2915. Voltage 1150 -> 1125. Ran the game for 80 mins, didn't crash. Real great that I bought a flagship card just to tune it down, I'm loving AMD. Anyways, I'll have to test it further tho. Attached is my current settings, hope it holds. Thanks for your idea anyways!
Thanks for the feedback! In short, the system holds up well at the base clocks as advertised. The boost feature works but might be a bit too strong for this particular chip. It could also be a sign of a bigger problem Nvidia faced with memory clocks on their RTX 2xxx models—severe crashes and instability were common until people adjusted the memory settings. Overall, it’s not a major issue, and your card should still perform nicely.
Have you attempted to use DDU? I've experienced unusual problems each time I updated Adrenalin without first running DDU. https://www.wagnardsoft.com/content/Down...-DDU-18077 When DDU is active, select the option to run in safe mode with network. After rebooting, navigate to DDU settings and enable "prevent windows from downloading and installing drivers." Perform a clean install and restart. For the latest drivers: https://www.amd.com/en/support/downloads...0-xtx.html Remember to try with default configurations and avoid overclocking.
I believe the issue might be related to the driver itself. I played Ghost of Tusima with the Adrenaline version from 2023Aug, and it ran smoothly without any crashes. However, after the latest patch, I had to upgrade my Adrenaline because the game wouldn't run on my old driver. That reminded me of a previous problem. Despite running a GPU load test for 30 minutes with the new driver, nothing happened. It seems the GPU handled stress well, but it still gave in during games. Could this be an issue with my motherboard? My PSU was upgraded to 1300 watts recently, which shouldn't be the case. I ran Memtest and CPU load tests, both passed. I think I've moved past the frustration and just want to figure out what's really going on.
Try this approach. Disable Windows Update installation for DDU and ensure the latest adrenaline drivers are installed. These new drivers perform excellently for me. Typically I run DDU to get them working properly; otherwise, I encounter strange problems when using multiple monitors.
In my experience, some games utilize the GPU more intensely than benchmarks do. I faced the same issue in Dying Light 2 with an RTX 3090. During tests, performance remained stable, but in-game power draw would cause overcurrent protection after about ten minutes of playtime. This seems linked to spikes from the GPU or CPU, especially when ray tracing is active. In your situation, it appears the PSU isn’t the problem—there’s enough power available. It’s likely the GPU struggles with higher frequency demands, possibly due to core or memory limitations. Downclocking could improve stability, and adjusting settings to match the cards’ recommended base clocks might help achieve a smoother run.
I discovered a reliable setting for my card with no recent crashes. Here’s my current configuration.