AMD Ryzen 3 CPUFreq Governor Benchmarks On Linux 4.13

[droste]

No... It's just waste of time of us, the Phoronix readers, because CPUFreq governor benchmarks without the power consumption metrics are meaningless. CPUFreq governors are there to reduce power consumption while keeping performance high in the first place. What is the conclusion? Should we use performance governor everywhere singly because it's more performant?

Didn't waste my time. I don't care too much about the power consumption, but the fan noise! So performance is not an option. I'm using schedutil which keeps the fan quiet and this benchmark shows me that I don't loose too much performance. So thanks :-)

[Modu]

Didn't waste my time. I don't care too much about the power consumption, but the fan noise! So performance is not an option. I'm using schedutil which keeps the fan quiet and this benchmark shows me that I don't loose too much performance. So thanks :-)

So you measure the power consumption by the fan noise...

[Anty]

What about buying better cooler?

[droste]

So you measure the power consumption by the fan noise...

No, I measure fan noise by fan noise. I doubt the efficiency is the same over a range of temperaturs. So twice as hot, doesn't mean twice the power consumption. And no, better cooling is not an option as all cooling solutions that fit into a normal computer case make some kind of sound as soon as the CPU is in full power mode. When idling with schedutil all my case fans are off and the CPU one is spinning at ~600rpm with the cpu at ~50°C. Almost no sound at all.

[sdack]

tl; dr use conservative or schedutil

Bah, no. Learn to adjust the parameters!

The ondemand governor is still the best. It's default settings are however more suitable for a laptop than a desktop or a server. Once adjusted can one still get very good power-savings from it with an idle machine, while it also gives a high performance under load. But if one doesn't learn about it and only keeps comparing governors at their default settings then that's about as stupid as comparing the top speeds of cars only in third gear.

Just add this to /etc/sysfs.conf when you have the sysfsutils installed:

# ondemand CPU governor, switch up at 50% load and stay up 10x longer
devices/system/cpu/cpufreq/ondemand/up_threshold = 50
devices/system/cpu/cpufreq/ondemand/sampling_down_factor = 10

[stefantalpalaru]

Thanks for the advice, sdack.

For non Debian distros, stick this in your "local" init script:

# ondemand CPU governor, switch up at 50% load and stay up 10x longer
echo 50 > /sys/devices/system/cpu/cpufreq/ondemand/up_threshold
echo 10 > /sys/devices/system/cpu/cpufreq/ondemand/sampling_down_factor

[kaseki]

On my Ryzen 7 1800X system with Mint 18.1 MATE 64-bit, with all eight cores' governors set at 'on-demand,' and with P-states active in my Asus C6H motherboard's BIOS, governors operating in the on-demand mode show operation within the range of P(0), P(1), and P(2). At present P(0) is set to 3.8 GHz. The others are at their defaults of 3.2 and 2.2, respectively. Whether the governor is controlling the BIOS or the governor is observing the BIOS in this mode is a matter of interest.

At present due to limitations with knowledge of the 8665 super-IO chip's interfaces, the governors are one of the few ways I have of observing what is going on when testing. debianxfce, how much of an otherwise fully loaded CPU's resources do you believe is used by its governor?

[MrCooper]

The ondemand governor is still the best. It's default settings are however more suitable for a laptop than a desktop or a server. Once adjusted can one still get very good power-savings from it with an idle machine, while it also gives a high performance under load. [...]

What's the advantage of tweaking ondemand over just using schedutil?

[sdack]

What's the advantage of tweaking ondemand over just using schedutil?

The ondemand governor gives you more power savings when a CPU is idle or nearly idle. It will leave the frequency at a low setting during those times and only when the load goes above the threshold will it switch up. schedutil bounces around too much for my taste and even when a machine seems idle will schedutil switch up the frequently for the smallest tasks, even when that's just the mail tool checking for new incoming mail for example.

With a tweaked sampling_down_factor of the ondemand governor will a CPU then also remain at a high frequency for longer even when the load briefly falls off. schedutil would here switch down only to switch up again. While schedutil is in theory the more dynamic method can a tweaked ondemand scheduler give you a better result all round. Too much switching of the frequency introduces micro delays, which isn't very desirable. Tweaking the ondemand scheduler reduces these transitions to as few as you actually need and can thereby give you stronger power-savings during idle times as well as a higher performance during high load times.

The real issue with the ondemand scheduler are its default settings. So will it at default only switch the frequency up when the load climbs above 95%. That's a rather high threshold, which many tasks often don't always reach or at least not consistently to also benefit from a CPU's maximum frequency. Thus lowering the threshold to a fair 50% puts it right into the middle of the scale. The sampling_down_factor determines how quickly it switches down again and when this gets increased does it lower the amount of frequency changes and it gives intense loads more of the high frequency you want them to have. In my case does a factor of 10 equal 0.1s or 100ms (default is 10ms), meaning, I get at most 10 frequency changes per second and not 100.

schedutil is a good all-rounder, which works best when you have frequently changing loads, but with little idle times and also few maximum load times. Personally is this not what I experience. My computers are often idle (0%-2% load) or they run lots of jobs (95-100%). So it's much more black&white for me. Hence do I find the ondemand scheduler to be the best, because when tweaked does it fit my needs the best.

[MrCooper]

The ondemand governor gives you more power savings when a CPU is idle or nearly idle. It will leave the frequency at a low setting during those times and only when the load goes above the threshold will it switch up. schedutil bounces around too much for my taste and even when a machine seems idle will schedutil switch up the frequently for the smallest tasks, even when that's just the mail tool checking for new incoming mail for example.

With a tweaked sampling_down_factor of the ondemand governor will a CPU then also remain at a high frequency for longer even when the load briefly falls off.

Maybe the defaults have changed since I stopped using ondemand on this laptop, but when I was using it, it would actually ramp up the clocks and hold them high for too long while the CPU was mostly idle, resulting in high temperature and fan noise. So I switched to conservative, which kept the temperature and fan noise in check, but took too long to ramp up the clocks for my taste. As soon as schedutil appeared, I switched to that, which now gives me the responsiveness I used to have with ondemand with the low temperature / fan noise I used to have with conservative.

Given my experience with idle temperature / fan noise, I find it hard to believe the "The ondemand governor gives you more power savings when a CPU is idle or nearly idle" claim without numbers backing it up.

Too much switching of the frequency introduces micro delays, which isn't very desirable.
[...]
In my case does a factor of 10 equal 0.1s or 100ms (default is 10ms), meaning, I get at most 10 frequency changes per second and not 100.

Have you measured any significant difference between the two, e.g. WRT throughput or power consumption with some reproducible workload?