Some Distributions Are Already Making Changes To Linux's Scheduler
Already it's looking like the research from the recently covered The Linux Scheduler: a Decade of Wasted Cores that called out the Linux kernel in being a poor scheduler is having an impact.
While we haven't seen any major upstream improvements yet to the Linux kernel scheduler, it looks like some parties are beginning to take note and better analyze the scheduler for possible performance improvements. Intel's Clear Linux distribution that's known for being focused on high out-of-the-box performance (even faster Intel graphics) already took action and landed some changes to their kernel's scheduler.
With yesterday's Clear Linux 7490 there was tweaking to their kernel scheduler as a result of that information, but I have yet to find out precisely what they changed with their scheduler. Nevertheless, I certainly plan on running some extra benchmarks to see the impact of their latest downstream kernel changes. We continue running daily benchmarks of Clear Linux via LinuxBenchmarking.com but at least with that test subset didn't see any performance changes, so looks like we'll need to increase the test coverage and I'm firing up some fresh installs on some other hardware to look for scheduler performance differences. Stay tuned or you can try out the Linux distribution for yourself at ClearLinux.org.
While we haven't seen any major upstream improvements yet to the Linux kernel scheduler, it looks like some parties are beginning to take note and better analyze the scheduler for possible performance improvements. Intel's Clear Linux distribution that's known for being focused on high out-of-the-box performance (even faster Intel graphics) already took action and landed some changes to their kernel's scheduler.
With yesterday's Clear Linux 7490 there was tweaking to their kernel scheduler as a result of that information, but I have yet to find out precisely what they changed with their scheduler. Nevertheless, I certainly plan on running some extra benchmarks to see the impact of their latest downstream kernel changes. We continue running daily benchmarks of Clear Linux via LinuxBenchmarking.com but at least with that test subset didn't see any performance changes, so looks like we'll need to increase the test coverage and I'm firing up some fresh installs on some other hardware to look for scheduler performance differences. Stay tuned or you can try out the Linux distribution for yourself at ClearLinux.org.
16 Comments