Linux Kernel Patch Could Revitalize Gaming on Older PCs

If you're running Linux on aging hardware and struggling with stuttering or low frame rates in games, a new kernel patch series might offer a lifeline. Developer Peter Zijlstra has posted a patch set called "sched: Flatten the pick" that aims to fix long-standing inefficiencies in how the Linux scheduler handles cgroup weight distribution across many CPU cores. The problem becomes especially noticeable on older systems with many cores or when multiple background tasks compete for resources. Below, we break down what this patch does, how it was tested, and what it means for your gaming experience.

What exactly is the scheduling problem this patch addresses?

The issue lies in how the Linux scheduler assigns CPU time to task groups (cgroups). Each group has a weight that determines its priority relative to others. On multi-core systems, the scheduler distributes that weight across every CPU, which fragments it. For instance, on a 64-core machine, a cgroup's weight per CPU drops to roughly that of a nice 19 task—very low priority. On 256-core servers, it becomes even smaller. When a game and background processes like Discord or a browser compete, the game ends up starved for CPU time, resulting in stutter and low frame rates. Peter Zijlstra called this fragmentation "a pain in the ar*e" because the math falls apart completely when all load lands on a single core.

How does the "Flatten the pick" patch fix this?

The patch addresses two underlying flaws. First, it corrects the weight inflation formula that normally multiplies group weight by the number of CPUs—this works when load is spread evenly, but if all tasks run on one core, the weight balloons far beyond nice -20, breaking scheduling decisions. Second, it restructures how the scheduler picks the next task. Currently, the kernel steps through multiple cgroup levels to find a task; the patch collapses all those levels into a single one, drastically reducing overhead. The result is a fairer, more efficient distribution of CPU time without the per-CPU weight fragmentation.

Who developed this patch and is it already in the mainline kernel?

The patch series was created by Peter Zijlstra, a prominent Linux kernel developer. He posted the second version (v2) under the title "sched: Flatten the pick." However, it has not yet been merged into the mainline kernel. The series still requires review and approval from the relevant kernel maintainers. It is likely to undergo several revisions before being accepted into a future kernel release. Gamers interested in immediate use would need to patch their own kernel or wait for it to be included in a distribution update.

What hardware and game did Peter Zijlstra use in his test?

To demonstrate the impact, Zijlstra used an Intel Core i7-2600K (a Sandy Bridge CPU from 2011) paired with an AMD Radeon RX 580 GPU—hardware that is more than a decade old for the CPU. He ran Shadows: Awakening from GOG via Lutris, using GE-Proton10-34 and Steam Runtime 3 (sniper). To simulate real-world background load, he launched 8 spinner processes (one per CPU thread) alongside the game. This setup intentionally stressed the scheduler to reproduce the stuttering and low FPS that many gamers with similar hardware experience.

What were the concrete performance improvements in the test?

The test compared the game running with the default scheduler time slice versus a slice shortened to one-tenth via the chrt command. The results, captured with MangoHud, were dramatic:

• Minimum FPS jumped from 3.8 to 20.6—a 5x improvement.
• Average FPS increased from 48.0 to 57.2.
• Maximum frame time dropped from 107.4 ms to 37.2 ms—eliminating most stutter.

While Zijlstra noted he hasn't compared against a kernel without the flat scheduling enabled, the shorter slice clearly helps the game breathe even under heavy background competition.

Why should this matter to gamers running modern PCs with many cores?

Although the test used an old quad-core CPU, the weight fragmentation problem grows worse with higher core counts. On a server-class 256-core system, the per-CPU priority of a cgroup is nearly negligible. For desktop gamers with 16, 32, or more cores, background tasks like Discord, a web browser with many tabs, or a system update can easily push a game into similar territory. The patch would benefit any multi-core setup where the scheduler must balance multiple cgroups. In essence, if this lands, both old and new hardware could see smoother gaming.

What can I do to get this improvement on my Linux system right now?

As the patch is not yet in the mainline kernel, you have a few options. Advanced users can apply the patch manually from the kernel mailing list and recompile their kernel—but this requires familiarity with kernel building. Alternatively, watch for distributions like Arch Linux or Fedora that often backport experimental patches more quickly. For most, the safest bet is to wait for the patch to be reviewed, revised, and merged into an official kernel release (likely 6.x or later). Stay tuned to sites like Phoronix for updates, or subscribe to the linux-kernel mailing list.