CPU Governors explained
Thanks to deedii for posting this in another forum:
Android CPU governors explained
What is a governor?
A governor is a driver for the regulation of CPUFreq - CPU frequency. As the name suggests, we, the Governor of the decision, when at full capacity, the MaxFreq - will be achieved or how fast the minFreq - - maximum frequency is reached minimum frequency or center frequency. He decides when, how and how long the CPU and still responds battery saving is still soft and still works.
There are many types of governors. Some are for single-core processors and some designed for dual-core processors. In stock kernel, there are five governors and quasar kernel, there are a lot more.
7: Min Max
1: OnDemand Governor:
This governor has a hair trigger for boosting clockspeed to the maximum speed set by the user. If the CPU load placed by the user abates, the OnDemand governor will slowly step back down through the kernel's frequency steppings until it settles at the lowest possible frequency, or the user executes another task to demand a ramp.
OnDemand has excellent interface fluidity because of its high-frequency bias, but it can also have a relatively negative effect on battery life versus other governors. OnDemand is commonly chosen by smartphone manufacturers because it is well-tested, reliable, and virtually guarantees the smoothest possible performance for the phone. This is so because users are vastly more likely to bitch about performance than they are the few hours of extra battery life another governor could have granted them.
This final fact is important to know before you read about the Interactive governor: OnDemand scales its clockspeed in a work queue context. In other words, once the task that triggered the clockspeed ramp is finished, OnDemand will attempt to move the clockspeed back to minimum. If the user executes another task that triggers OnDemand's ramp, the clockspeed will bounce from minimum to maximum. This can happen especially frequently if the user is multi-tasking. This, too, has negative implications for battery life.
Basically an ondemand with suspend/wake profiles. This governor is supposed to be a battery friendly ondemand. When screen is off, max frequency is capped at 500 mhz. Even though ondemand is the default governor in many kernel and is considered safe/stable, the support for ondemand/ondemandX depends on CPU capability to do fast frequency switching which are very low latency frequency transitions. I have read somewhere that the performance of ondemand/ondemandx were significantly varying for different i/o schedulers. This is not true for most of the other governors. I personally feel ondemand/ondemandx goes best with SIO I/O scheduler.
3: Performance Governor:
This locks the phone's CPU at maximum frequency. While this may sound like an ugly idea, there is growing evidence to suggest that running a phone at its maximum frequency at all times will allow a faster race-to-idle. Race-to-idle is the process by which a phone completes a given task, such as syncing email, and returns the CPU to the extremely efficient low-power state. This still requires extensive testing, and a kernel that properly implements a given CPU's C-states (low power states).
4: Powersave Governor:
The opposite of the Performance governor, the Powersave governor locks the CPU frequency at the lowest frequency set by the user.
This biases the phone to prefer the lowest possible clockspeed as often as possible. In other words, a larger and more persistent load must be placed on the CPU before the conservative governor will be prompted to raise the CPU clockspeed. Depending on how the developer has implemented this governor, and the minimum clockspeed chosen by the user, the conservative governor can introduce choppy performance. On the other hand, it can be good for battery life.
The Conservative Governor is also frequently described as a "slow OnDemand," if that helps to give you a more complete picture of its functionality.
6: Userspace Governor:
This governor, exceptionally rare for the world of mobile devices, allows any program executed by the user to set the CPU's operating frequency. This governor is more common amongst servers or desktop PCs where an application (like a power profile app) needs privileges to set the CPU clockspeed.
7: Min Max
well this governor makes use of only min & maximum frequency based on workload... no intermediate frequencies are used.
8: Interactive Governor:
Much like the OnDemand governor, the Interactive governor dynamically scales CPU clockspeed in response to the workload placed on the CPU by the user. This is where the similarities end. Interactive is significantly more responsive than OnDemand, because it's faster at scaling to maximum frequency.
Unlike OnDemand, which you'll recall scales clockspeed in the context of a work queue, Interactive scales the clockspeed over the course of a timer set arbitrarily by the kernel developer. In other words, if an application demands a ramp to maximum clockspeed (by placing 100% load on the CPU), a user can execute another task before the governor starts reducing CPU frequency. This can eliminate the frequency bouncing discussed in the OnDemand section. Because of this timer, Interactive is also better prepared to utilize intermediate clockspeeds that fall between the minimum and maximum CPU frequencies. This is another pro-battery life benefit of Interactive.
However, because Interactive is permitted to spend more time at maximum frequency than OnDemand (for device performance reasons), the battery-saving benefits discussed above are effectively negated. Long story short, Interactive offers better performance than OnDemand (some say the best performance of any governor) and negligibly different battery life.
Interactive also makes the assumption that a user turning the screen on will shortly be followed by the user interacting with some application on their device. Because of this, screen on triggers a ramp to maximum clockspeed, followed by the timer behavior described above.
9: InteractiveX Governor:
Created by kernel developer "Imoseyon," the InteractiveX governor is based heavily on the Interactive governor, enhanced with tuned timer parameters to better balance battery vs. performance. The InteractiveX governor's defining feature, however, is that it locks the CPU frequency to the user's lowest defined speed when the screen is off.
Is based on the concept of the interactive governor.
I have always agreed that in theory the way interactive works – by taking over the idle loop – is very attractive. I have never managed to tweak it so it would behave decently in real life. Smartass is a complete rewrite of the code plus more. I think its a success. Performance is on par with the “old” minmax and I think smartass is a bit more responsive. Battery life is hard to quantify precisely but it does spend much more time at the lower frequencies.
Smartass will also cap the max frequency when sleeping to 352Mhz (or if your min frequency is higher than 352 – why?! – it will cap it to your min frequency). Lets take for example the 528/176 kernel, it will sleep at 352/176. No need for sleep profiles any more!"
Version 2 of the original smartass governor from Erasmux. Another favorite for many a people. The governor aim for an "ideal frequency", and ramp up more aggressively towards this freq and less aggressive after. It uses different ideal frequencies for screen on and screen off, namely awake_ideal_freq and sleep_ideal_freq. This governor scales down CPU very fast (to hit sleep_ideal_freq soon) while screen is off and scales up rapidly to awake_ideal_freq (500 mhz for GS2 by default) when screen is on. There's no upper limit for frequency while screen is off (unlike Smartass). So the entire frequency range is available for the governor to use during screen-on and screen-off state. The motto of this governor is a balance between performance and battery.
A new governor wrote based on conservative with some smartass features, it scales accordingly to conservatives laws. So it will start from the bottom, take a load sample, if it's above the upthreshold, ramp up only one speed at a time, and ramp down one at a time. It will automatically cap the off screen speeds to 245Mhz, and if your min freq is higher than 245mhz, it will reset the min to 120mhz while screen is off and restore it upon screen awakening, and still scale accordingly to conservatives laws. So it spends most of its time at lower frequencies. The goal of this is to get the best battery life with decent performance. It will give the same performance as conservative right now, it will get tweaked over time.
Lagfree is similar to ondemand. Main difference is it's optimization to become more battery friendly. Frequency is gracefully decreased and increased, unlike ondemand which jumps to 100% too often. Lagfree does not skip any frequency step while scaling up or down. Remember that if there's a requirement for sudden burst of power, lagfree can not satisfy that since it has to raise cpu through each higher frequency step from current. Some users report that video playback using lagfree stutters a little.
The same as the Smartass “governor” But MUCH more aggressive & across the board this one has a better battery life that is about a third better than stock KERNEL
Similar to smartassV2. More aggressive ramping, so more performance, less battery
Another smartassV2 based governor. Achieves good balance between performance & battery as compared to brazilianwax.
This governor from Ezekeel is basically an ondemand with an additional parameter min_time_state to specify the minimum time CPU stays on a frequency before scaling up/down. The Idea here is to eliminate any instabilities caused by fast frequency switching by ondemand. Lazy governor polls more often than ondemand, but changes frequency only after completing min_time_state on a step overriding sampling interval. Lazy also has a screenoff_maxfreq parameter which when enabled will cause the governor to always select the maximum frequency while the screen is off.
Lionheart is a conservative-based governor which is based on samsung's update3 source.
The tunables (such as the thresholds and sampling rate) were changed so the governor behaves more like the performance one, at the cost of battery as the scaling is very aggressive.
LionheartX is based on Lionheart but has a few changes on the tunables and features a suspend profile based on Smartass governor.
Intellidemand aka Intelligent Ondemand from Faux is yet another governor that's based on ondemand. Unlike what some users believe, this governor is not the replacement for OC Daemon (Having different governors for sleep and awake). The original intellidemand behaves differently according to GPU usage. When GPU is really busy (gaming, maps, benchmarking, etc) intellidemand behaves like ondemand. When GPU is 'idling' (or moderately busy), intellidemand limits max frequency to a step depending on frequencies available in your device/kernel for saving battery. This is called browsing mode. We can see some 'traces' of interactive governor here. Frequency scale-up decision is made based on idling time of CPU. Lower idling time (<20%) causes CPU to scale-up from current frequency. Frequency scale-down happens at steps=5% of max frequency. (This parameter is tunable only in conservative, among the popular governors)
To sum up, this is an intelligent ondemand that enters browsing mode to limit max frequency when GPU is idling, and (exits browsing mode) behaves like ondemand when GPU is busy; to deliver performance for gaming and such. Intellidemand does not jump to highest frequency when screen is off.
21: Hotplug Governor:
The “hotplug” governor scales CPU frequency based on load, similar to “ondemand”. It scales up to the highest frequency when “up_threshold” is crossed and scales down one frequency at a time when “down_threshold” is crossed. Unlike those governors, target frequencies are determined by directly accessing the CPUfreq frequency table, instead of taking some percentage of maximum available frequency.
The key difference in the “hotplug” governor is that it will disable auxillary CPUs when the system is very idle, and enable them again once the system becomes busy. This is achieved by averaging load over multiple sampling periods; if CPUs were online or offlined based on a single sampling period then thrashing will occur.
Sysfs entries exist for “hotplug_in_sampling_periods” and for “hotplug_out_sampling_periods” which determine how many consecutive periods get averaged to determine if auxillery CPUs should be onlined or offlined. Defaults are 5 periods and 20 periods respectively. Otherwise the standard sysfs entries you might find for “ondemand” and “conservative” governors are there.
Obviously, this governor is only available on multi-core devices.
in short words this govenor is build on “ondemand” but increases the C4 state time of the CPU and doing so trying to save juice.
23: Basically interactive governor with added smartass bits and variable (as opposed to fixed amout) frequency scaling, based on currently occuring cpu loads. Has, like smartass, a sleep profile built-in. See link for details on exact scaling.
24: Abyssplug governor is a modified hotplug governor.
25. BadAss Governor:
Badass removes all of this "fast peaking" to the max frequency. On a typical system the cpu won't go above 918Mhz and therefore stay cool and will use less power. To trigger a frequency increase, the system must run a bit @ 918Mhz with high load, then the frequency is bumped to 1188Mhz. If that is still not enough the governor gives you full throttle. (this transition should not take longer than 1-2 seconds, depending on the load your system is experiencing)
Badass will also take the gpu load into consideration. If the gpu is moderately busy it will bypass the above check and clock the cpu with 1188Mhz. If the gpu is crushed under load, badass will lift the restrictions to the cpu.
Ondemand scales to the highest frequency as soon as a load occurs. Conservative scales upward based on the frequency step variable which means for the most part will scale through every frequency to achieve the target load thresholds. What this practically means is ondemand is prone to wasting power on unneeded clock cycles. Ondemand also features something called a down differential, this variable determines how long the governor will remain at the given frequency before scaling down. Conservative does not have this, but instead relies on having a down threshold which insures that as soon as the load drops below a given variable it scales down as fast as the sampling rate allows. The result to this is a governor which attempts to keep the load level tolerable and save you battery! Now ! Ktoonservative Is that but in addition contains a hotpluging variable which determines when the second core comes online. The governor shuts the core off when it returns to the second lowest frequency thus giving us a handle on the second performance factor in our CPUs behavior. While by default conservative is a poor performer it can be made to perform comparably to even performance governor. Here are some settings to discuss and start with. They are slightly less battery friendly under a load but very very well performing.
So far, all I have found about this Governor is that it belongs in the interactive family. I'll update this when I find more
The Sleepy (formerly known as Solo) is an attempt to strike a balance between performance and battery power to create. It is based on the getweakten Ondemand of Arighi and is optimized for the SGS2. It may include imoseyon's Ondemandx with some tweaks Down_sampling and other features that set by the user through the sysfs of "echo" call. Sleepy is the behavior of Ondemandx when he is in action, very similar.
The Hyper (formerly known as kenobi) is an aggressive smart and smooth, optimized for SGS2 getweakt and, based on the Ondemand, which was getweakt of Arighi and was equipped with several features of Ondemandx suspend imoseyon. (Added by sysfs, the settings suspend_freq and suspend Imoseyon's code) is the behavior of the hyper Ondemand if he is in action, very similar. He also has the Arighi's fast_start deep_sleep and detection features. In addition, the maximum frequency is in suspend mode 500Mhz.
Credits goes to:
What is a scheduler?
In a multitasking operating system, there must be an instance, the processes that want to run, CPU time and allocates it "goes to sleep" after the allotted time (timeslice) again. This instance is called the scheduler, such as opening and closing applications. that is, how fast they are open and how long they are kept in RAM.
I / O scheduler can have many purposes like:
To minimize time searching on the hard disk
Set priorities for specific process requests
To regulate a particular portion of the bandwidth of the data carrier to each running process
To guarantee certain process requests within a certain time
Which scheduler are available?
Two important things here are indicative of that event:
- Looking on the flash drive is very slow from Equip
- Write operations while at any time are processed, however, be read operations preferred, ie, this scheduler returns the read operations a higher priority than the write operations.
- Requests of read accesses are never treated secondarily, that has equally good reading performance on flash drives like the noop
- Requests from process operations are not always available
- Reduced write performance on high-performance hard drives
The CFQ - Completely Fair Queuing - similar to the Dead Line maintains a scalable continuous Prozess-I/O-Warteschlange, ie the available I / O bandwidth tried fairly and evenly to all I / O requests to distribute. He created a statistics between blocks and processes. With these statistics it can "guess" when the next block is requested by what process, ie each process queue contains requests of synchronous processes, which in turn is dependent upon the priority of the original process. There is a V2 and the CFQ has some fixes, such as were the I / O request, hunger, and some small search backward integrated to improve the responsiveness.
- Has the goal of a balanced I / O performance to deliver
- The easiest way to set
- Excellent on multiprocessor systems
- Best performance of the database after the deadline
- Some reported user that the media scanning would take this very very long time and this by the very fair and even distribution of bandwidth on the I / O operations during the boot process is conditioned with the media scanning is not necessarily the highest should have priority
- Jitter (worst case delay) can sometimes be very high because the number of competing with each other process tasks
This scheduler has the goal of reducing I / O wait time of a process of inquiry. This is done using the block numbers of the data on the drive. This also blocks an outlying block numbers are processed, each request receives a maximum delivery time. This is in addition to the Governor BFQ very popular and in many well known kernels, such as the Nexus S Netarchy. He was indeed better than the BFQ, but compared to the VR he will be weaker.
- Is nearly a real-time scheduler.
- Characterized by reducing the waiting time of each process from - best scheduler for database access and queries.
- Bandwidth requirements of a process, eg what percentage does a CPU is easy to calculate.
- As the Governor-noop ideal for flash drives
- If the system is overloaded, can go a lost set of processes, and is not as easy to predict
It aims to achieve with minimal effort at a low latency I / O requests. Not a priority to put in queue, instead simply merge the requests. This scheduler is a mix between the noop and deadline. With him there is no conversion or sorting of requests.
- It is simple and stable. - Minimized Starvations (starvation) for inquiries
- Slow random write speeds on flash drives as opposed to other schedulers. - Sequential read speeds on flash drives, not as good
The noop scheduler is the simplest of them. He is best suited for storage devices that are not subject to mechanical movements, such as our flash drives in our SGSII's to use to access the data. The advantage is that flash drives do not require rearrangement of the I / O requests, unlike normal hard drives. ie the data that come first are written first. He's basically not a real scheduler, as it leaves the scheduling of the hardware.
- Adds all incoming I / O requests in a first-come-who-first-served queue and implements requests with the fewest number of CPU cycles, so also battery friendly
- Is suitable for flash drives because there is no search errors
- Good data throughput on db systems
- Reducing the number of CPU cycles corresponds to a simultaneous decline in performance einhergehendem
Unlike other scheduling software, synchronous and asynchronous requests are not handled separately, but it will impose a fair and balanced within this deadline requests, that the next request to be served is a function of distance from the last request. The VR is a very good scheduler with elements of the deadline scheduler. He will probably be the best for MTD Android devices. He is the one who can make the most of the benchmark points, but he is also an unstable schedulers, because his performance falter. Sometimes they fluctuate below the average, sometimes it fluctuates above the average, but if above, then he is the best.
- Is the best scheduler for benchmarks
- Performance variability can lead to different results
- Very often unstable or unzverlässig
As the name suggests, it is more of a simple or simple scheduler. Especially suitable for EMMC devices. He is reliable, maybe not as good as the VR, when this time has a good day, but he is despite all this very performance-based and does his best. At the moment it is the default scheduler in quasar kernel.
Advantages: - not known
Cons: - not known
Instead requests divided into time segments as the CFQ has, on the BFQ budget. The flash drive will be granted an active process until it has exhausted its budget (number of sectors on the flash drive). The awards BFQ high budget does not read tasks.
- Has a very good USB data transfer rate.
- Be the best scheduler for playback of HD video recording and video streaming (due to less jitter than CFQ Scheduler, and others)
- Regarded as a very precise working Scheduler
- Delivers 30% more throughput than CFQ
- Not the best scheduler for benchmarks - higher budgets that were allocated to a process that can affect the interactivity and bring with it increased latency.
How can I change the governor and scheduler?
There are two ways to change the governor and schedulers, as well as the settings for the Governorn. Either manually, in which you use a file manager like Root Explorer and then knows how to / sys / devices / system and then change the files to his wishes, provided you what you're doing, or via a graphical interface or by phone as SetCPU Voltage Control. These are the most prominent apps when it comes to adjusting the governor and / or scheduler.
- SetCPU are, besides the possibility of the clock speed of the CPU, setting profiles in certain situations, only to change the way the governor. The scheduler can not change it.
- Voltage control can alter both the governor and the scheduler, but has no way to adjust behavior profiles. While you can set various overclocking, Governor and scheduler profiles manually, but nothing more. Nevertheless, I prefer the VC, since it is simple and gives me the opportunity to change the scheduler.
Credit goes to Tinzdroid