Ryzen 7 2700X CPUFreq Scaling Governor Benchmarks On Ubuntu Linux
With this week's Ryzen 5 2600X + Ryzen 7 2700X benchmarks some thought the CPUFreq scaling driver or rather its governors may have been limiting the performance of these Zen+ CPUs, so I ran some additional benchmarks this weekend.
Those launch-day Ryzen 5 2600X / Ryzen 7 2700X Ubuntu Linux benchmarks were using the "performance" governor, but some have alleged that the performance governor may now actually hurt AMD systems... Ondemand, of course, is the default CPUFreq governor on Ubuntu and most other Linux distributions. Some also have said the "schedutil" governor that makes use of the kernel's scheduler utilization data may do better on AMD. So I ran some extra benchmarks while changing between CPUFreq's ondemand (default), performance (normally the best for performance, and what was used in our CPU tests), schedutil (the newest option), and powersave (if you really just care about conserving power).
Tests were done via the Phoronix Test Suite and OpenBenchmarking.org.
In the few graphical tests, the performance governor performed the same as ondemand and schedutil... While obviously powersave led to the slowest performance.
And in the CPU tests, performance tended to be the best or at worst was in line with ondemand and schedutil.
I did not find any situation where using the performance governor led to worse performance.
Nor did Schedutil end up being a magic bullet.
More AMD Ryzen 7 2700X Linux benchmarks via this OpenBenchmarking.org result file.
Those launch-day Ryzen 5 2600X / Ryzen 7 2700X Ubuntu Linux benchmarks were using the "performance" governor, but some have alleged that the performance governor may now actually hurt AMD systems... Ondemand, of course, is the default CPUFreq governor on Ubuntu and most other Linux distributions. Some also have said the "schedutil" governor that makes use of the kernel's scheduler utilization data may do better on AMD. So I ran some extra benchmarks while changing between CPUFreq's ondemand (default), performance (normally the best for performance, and what was used in our CPU tests), schedutil (the newest option), and powersave (if you really just care about conserving power).
Tests were done via the Phoronix Test Suite and OpenBenchmarking.org.
In the few graphical tests, the performance governor performed the same as ondemand and schedutil... While obviously powersave led to the slowest performance.
And in the CPU tests, performance tended to be the best or at worst was in line with ondemand and schedutil.
I did not find any situation where using the performance governor led to worse performance.
Nor did Schedutil end up being a magic bullet.
More AMD Ryzen 7 2700X Linux benchmarks via this OpenBenchmarking.org result file.
24 Comments