AMD Confirms Radeon Driver Bug: atieclxx.exe Writes to SSD Around the Clock

⚡ The 5-Second Summary

• The atieclxx.exe process from Radeon drivers continuously writes to disk: about 1 MB/s, and up to 5 MB/s for some users.
• AMD has acknowledged the problem: the AMD External Events Utility service conflicts with Microsoft's PCIe filter driver.
• The fix is in final testing and will be included in one of the upcoming Radeon driver releases.

🔍 What Happened

The culprit is the AMD External Events Client Module, also known as the AMD External Events Utility, also known as the atieclxx.exe process. It's a Windows service that monitors the display: it detects when a new monitor is connected, handles screen hotkeys, and keeps AMD FreeSync running. Now, on top of all that, it's hammering the disk — constantly and without pauses.

On an SSD, these writes are barely noticeable — the drive is fast, and the counter grows silently. About 1 MB/s in the background — for hours on end. For some users, the counter ramps up to 5 MB/s. For a cheap 128 GB QLC drive with a low TBW rating, that's bad news: the rewrite endurance runs out faster than the manufacturer promised.

The bug isn't new. It was fixed in AMD chipset drivers in version 8.09.28.1124, but it resurfaced in the latest graphics driver — as reported by Habr. An AMD representative on Reddit confirmed that the service conflicts with Microsoft's PCIe filter driver, and that disabling ReBAR in the BIOS works around the problem.

🧠 Our Take

Why this surfaced right now is a question about driver architecture. AMD ships the same code between chipset and graphics packages. The patch was applied in the chipset branch, but in the GPU branch, in our assessment, it simply wasn't carried over. A classic case of two teams building drivers from shared modules falling out of sync.

What this means in practice. Formally, 1 MB/s is trivial — any browser writes more. But these writes never stop: the user isn't doing anything, yet the drive keeps working. On a budget laptop with a QLC drive and a couple of years of warranty, that background activity eats into endurance noticeably. We believe the main risk here isn't the speed but the invisibility: an ordinary user won't dig into the resource monitor looking for atieclxx.exe.

The winners are owners of expensive SSDs with high TBW — they won't notice anything. The losers are those who bought a budget drive and are also running the latest Adrenalin. It's especially awkward for AMD: the company has spent recent months promising to fix Radeon bugs quickly, yet here it had to explain itself in Reddit comments while the fix is still being tested.

⚠️ What We Don't Know

The source doesn't specify which graphics driver version introduced the bug — only the chipset fix 8.09.28.1124 and a mention of version 26.3.1 as the "previous" version for file replacement are given. It's unclear how many people actually hit 5 MB/s and on what configurations. The advice to copy amdkmpfd.ctz, amdkmpfd.itz, and amdkmpfd.stz into System32\AMD is a user discovery without verification from AMD. Fixing drivers by manually swapping files should only be done with a backup.