1.. SPDX-License-Identifier: GPL-2.0 2.. include:: <isonum.txt> 3 4.. |intel_pstate| replace:: :doc:`intel_pstate <intel_pstate>` 5 6======================= 7CPU Performance Scaling 8======================= 9 10:Copyright: |copy| 2017 Intel Corporation 11 12:Author: Rafael J. Wysocki <rafael.j.wysocki@intel.com> 13 14 15The Concept of CPU Performance Scaling 16====================================== 17 18The majority of modern processors are capable of operating in a number of 19different clock frequency and voltage configurations, often referred to as 20Operating Performance Points or P-states (in ACPI terminology). As a rule, 21the higher the clock frequency and the higher the voltage, the more instructions 22can be retired by the CPU over a unit of time, but also the higher the clock 23frequency and the higher the voltage, the more energy is consumed over a unit of 24time (or the more power is drawn) by the CPU in the given P-state. Therefore 25there is a natural tradeoff between the CPU capacity (the number of instructions 26that can be executed over a unit of time) and the power drawn by the CPU. 27 28In some situations it is desirable or even necessary to run the program as fast 29as possible and then there is no reason to use any P-states different from the 30highest one (i.e. the highest-performance frequency/voltage configuration 31available). In some other cases, however, it may not be necessary to execute 32instructions so quickly and maintaining the highest available CPU capacity for a 33relatively long time without utilizing it entirely may be regarded as wasteful. 34It also may not be physically possible to maintain maximum CPU capacity for too 35long for thermal or power supply capacity reasons or similar. To cover those 36cases, there are hardware interfaces allowing CPUs to be switched between 37different frequency/voltage configurations or (in the ACPI terminology) to be 38put into different P-states. 39 40Typically, they are used along with algorithms to estimate the required CPU 41capacity, so as to decide which P-states to put the CPUs into. Of course, since 42the utilization of the system generally changes over time, that has to be done 43repeatedly on a regular basis. The activity by which this happens is referred 44to as CPU performance scaling or CPU frequency scaling (because it involves 45adjusting the CPU clock frequency). 46 47 48CPU Performance Scaling in Linux 49================================ 50 51The Linux kernel supports CPU performance scaling by means of the ``CPUFreq`` 52(CPU Frequency scaling) subsystem that consists of three layers of code: the 53core, scaling governors and scaling drivers. 54 55The ``CPUFreq`` core provides the common code infrastructure and user space 56interfaces for all platforms that support CPU performance scaling. It defines 57the basic framework in which the other components operate. 58 59Scaling governors implement algorithms to estimate the required CPU capacity. 60As a rule, each governor implements one, possibly parametrized, scaling 61algorithm. 62 63Scaling drivers talk to the hardware. They provide scaling governors with 64information on the available P-states (or P-state ranges in some cases) and 65access platform-specific hardware interfaces to change CPU P-states as requested 66by scaling governors. 67 68In principle, all available scaling governors can be used with every scaling 69driver. That design is based on the observation that the information used by 70performance scaling algorithms for P-state selection can be represented in a 71platform-independent form in the majority of cases, so it should be possible 72to use the same performance scaling algorithm implemented in exactly the same 73way regardless of which scaling driver is used. Consequently, the same set of 74scaling governors should be suitable for every supported platform. 75 76However, that observation may not hold for performance scaling algorithms 77based on information provided by the hardware itself, for example through 78feedback registers, as that information is typically specific to the hardware 79interface it comes from and may not be easily represented in an abstract, 80platform-independent way. For this reason, ``CPUFreq`` allows scaling drivers 81to bypass the governor layer and implement their own performance scaling 82algorithms. That is done by the |intel_pstate| scaling driver. 83 84 85``CPUFreq`` Policy Objects 86========================== 87 88In some cases the hardware interface for P-state control is shared by multiple 89CPUs. That is, for example, the same register (or set of registers) is used to 90control the P-state of multiple CPUs at the same time and writing to it affects 91all of those CPUs simultaneously. 92 93Sets of CPUs sharing hardware P-state control interfaces are represented by 94``CPUFreq`` as struct cpufreq_policy objects. For consistency, 95struct cpufreq_policy is also used when there is only one CPU in the given 96set. 97 98The ``CPUFreq`` core maintains a pointer to a struct cpufreq_policy object for 99every CPU in the system, including CPUs that are currently offline. If multiple 100CPUs share the same hardware P-state control interface, all of the pointers 101corresponding to them point to the same struct cpufreq_policy object. 102 103``CPUFreq`` uses struct cpufreq_policy as its basic data type and the design 104of its user space interface is based on the policy concept. 105 106 107CPU Initialization 108================== 109 110First of all, a scaling driver has to be registered for ``CPUFreq`` to work. 111It is only possible to register one scaling driver at a time, so the scaling 112driver is expected to be able to handle all CPUs in the system. 113 114The scaling driver may be registered before or after CPU registration. If 115CPUs are registered earlier, the driver core invokes the ``CPUFreq`` core to 116take a note of all of the already registered CPUs during the registration of the 117scaling driver. In turn, if any CPUs are registered after the registration of 118the scaling driver, the ``CPUFreq`` core will be invoked to take note of them 119at their registration time. 120 121In any case, the ``CPUFreq`` core is invoked to take note of any logical CPU it 122has not seen so far as soon as it is ready to handle that CPU. [Note that the 123logical CPU may be a physical single-core processor, or a single core in a 124multicore processor, or a hardware thread in a physical processor or processor 125core. In what follows "CPU" always means "logical CPU" unless explicitly stated 126otherwise and the word "processor" is used to refer to the physical part 127possibly including multiple logical CPUs.] 128 129Once invoked, the ``CPUFreq`` core checks if the policy pointer is already set 130for the given CPU and if so, it skips the policy object creation. Otherwise, 131a new policy object is created and initialized, which involves the creation of 132a new policy directory in ``sysfs``, and the policy pointer corresponding to 133the given CPU is set to the new policy object's address in memory. 134 135Next, the scaling driver's ``->init()`` callback is invoked with the policy 136pointer of the new CPU passed to it as the argument. That callback is expected 137to initialize the performance scaling hardware interface for the given CPU (or, 138more precisely, for the set of CPUs sharing the hardware interface it belongs 139to, represented by its policy object) and, if the policy object it has been 140called for is new, to set parameters of the policy, like the minimum and maximum 141frequencies supported by the hardware, the table of available frequencies (if 142the set of supported P-states is not a continuous range), and the mask of CPUs 143that belong to the same policy (including both online and offline CPUs). That 144mask is then used by the core to populate the policy pointers for all of the 145CPUs in it. 146 147The next major initialization step for a new policy object is to attach a 148scaling governor to it (to begin with, that is the default scaling governor 149determined by the kernel command line or configuration, but it may be changed 150later via ``sysfs``). First, a pointer to the new policy object is passed to 151the governor's ``->init()`` callback which is expected to initialize all of the 152data structures necessary to handle the given policy and, possibly, to add 153a governor ``sysfs`` interface to it. Next, the governor is started by 154invoking its ``->start()`` callback. 155 156That callback is expected to register per-CPU utilization update callbacks for 157all of the online CPUs belonging to the given policy with the CPU scheduler. 158The utilization update callbacks will be invoked by the CPU scheduler on 159important events, like task enqueue and dequeue, on every iteration of the 160scheduler tick or generally whenever the CPU utilization may change (from the 161scheduler's perspective). They are expected to carry out computations needed 162to determine the P-state to use for the given policy going forward and to 163invoke the scaling driver to make changes to the hardware in accordance with 164the P-state selection. The scaling driver may be invoked directly from 165scheduler context or asynchronously, via a kernel thread or workqueue, depending 166on the configuration and capabilities of the scaling driver and the governor. 167 168Similar steps are taken for policy objects that are not new, but were "inactive" 169previously, meaning that all of the CPUs belonging to them were offline. The 170only practical difference in that case is that the ``CPUFreq`` core will attempt 171to use the scaling governor previously used with the policy that became 172"inactive" (and is re-initialized now) instead of the default governor. 173 174In turn, if a previously offline CPU is being brought back online, but some 175other CPUs sharing the policy object with it are online already, there is no 176need to re-initialize the policy object at all. In that case, it only is 177necessary to restart the scaling governor so that it can take the new online CPU 178into account. That is achieved by invoking the governor's ``->stop`` and 179``->start()`` callbacks, in this order, for the entire policy. 180 181As mentioned before, the |intel_pstate| scaling driver bypasses the scaling 182governor layer of ``CPUFreq`` and provides its own P-state selection algorithms. 183Consequently, if |intel_pstate| is used, scaling governors are not attached to 184new policy objects. Instead, the driver's ``->setpolicy()`` callback is invoked 185to register per-CPU utilization update callbacks for each policy. These 186callbacks are invoked by the CPU scheduler in the same way as for scaling 187governors, but in the |intel_pstate| case they both determine the P-state to 188use and change the hardware configuration accordingly in one go from scheduler 189context. 190 191The policy objects created during CPU initialization and other data structures 192associated with them are torn down when the scaling driver is unregistered 193(which happens when the kernel module containing it is unloaded, for example) or 194when the last CPU belonging to the given policy in unregistered. 195 196 197Policy Interface in ``sysfs`` 198============================= 199 200During the initialization of the kernel, the ``CPUFreq`` core creates a 201``sysfs`` directory (kobject) called ``cpufreq`` under 202:file:`/sys/devices/system/cpu/`. 203 204That directory contains a ``policyX`` subdirectory (where ``X`` represents an 205integer number) for every policy object maintained by the ``CPUFreq`` core. 206Each ``policyX`` directory is pointed to by ``cpufreq`` symbolic links 207under :file:`/sys/devices/system/cpu/cpuY/` (where ``Y`` represents an integer 208that may be different from the one represented by ``X``) for all of the CPUs 209associated with (or belonging to) the given policy. The ``policyX`` directories 210in :file:`/sys/devices/system/cpu/cpufreq` each contain policy-specific 211attributes (files) to control ``CPUFreq`` behavior for the corresponding policy 212objects (that is, for all of the CPUs associated with them). 213 214Some of those attributes are generic. They are created by the ``CPUFreq`` core 215and their behavior generally does not depend on what scaling driver is in use 216and what scaling governor is attached to the given policy. Some scaling drivers 217also add driver-specific attributes to the policy directories in ``sysfs`` to 218control policy-specific aspects of driver behavior. 219 220The generic attributes under :file:`/sys/devices/system/cpu/cpufreq/policyX/` 221are the following: 222 223``affected_cpus`` 224 List of online CPUs belonging to this policy (i.e. sharing the hardware 225 performance scaling interface represented by the ``policyX`` policy 226 object). 227 228``bios_limit`` 229 If the platform firmware (BIOS) tells the OS to apply an upper limit to 230 CPU frequencies, that limit will be reported through this attribute (if 231 present). 232 233 The existence of the limit may be a result of some (often unintentional) 234 BIOS settings, restrictions coming from a service processor or other 235 BIOS/HW-based mechanisms. 236 237 This does not cover ACPI thermal limitations which can be discovered 238 through a generic thermal driver. 239 240 This attribute is not present if the scaling driver in use does not 241 support it. 242 243``cpuinfo_cur_freq`` 244 Current frequency of the CPUs belonging to this policy as obtained from 245 the hardware (in KHz). 246 247 This is expected to be the frequency the hardware actually runs at. 248 If that frequency cannot be determined, this attribute should not 249 be present. 250 251``cpuinfo_avg_freq`` 252 An average frequency (in KHz) of all CPUs belonging to a given policy, 253 derived from a hardware provided feedback and reported on a time frame 254 spanning at most few milliseconds. 255 256 This is expected to be based on the frequency the hardware actually runs 257 at and, as such, might require specialised hardware support (such as AMU 258 extension on ARM). If one cannot be determined, this attribute should 259 not be present. 260 261 Note that failed attempt to retrieve current frequency for a given 262 CPU(s) will result in an appropriate error, i.e.: EAGAIN for CPU that 263 remains idle (raised on ARM). 264 265``cpuinfo_max_freq`` 266 Maximum possible operating frequency the CPUs belonging to this policy 267 can run at (in kHz). 268 269``cpuinfo_min_freq`` 270 Minimum possible operating frequency the CPUs belonging to this policy 271 can run at (in kHz). 272 273``cpuinfo_transition_latency`` 274 The time it takes to switch the CPUs belonging to this policy from one 275 P-state to another, in nanoseconds. 276 277 If unknown or if known to be so high that the scaling driver does not 278 work with the `ondemand`_ governor, -1 (:c:macro:`CPUFREQ_ETERNAL`) 279 will be returned by reads from this attribute. 280 281``related_cpus`` 282 List of all (online and offline) CPUs belonging to this policy. 283 284``scaling_available_frequencies`` 285 List of available frequencies of the CPUs belonging to this policy 286 (in kHz). 287 288``scaling_available_governors`` 289 List of ``CPUFreq`` scaling governors present in the kernel that can 290 be attached to this policy or (if the |intel_pstate| scaling driver is 291 in use) list of scaling algorithms provided by the driver that can be 292 applied to this policy. 293 294 [Note that some governors are modular and it may be necessary to load a 295 kernel module for the governor held by it to become available and be 296 listed by this attribute.] 297 298``scaling_cur_freq`` 299 Current frequency of all of the CPUs belonging to this policy (in kHz). 300 301 In the majority of cases, this is the frequency of the last P-state 302 requested by the scaling driver from the hardware using the scaling 303 interface provided by it, which may or may not reflect the frequency 304 the CPU is actually running at (due to hardware design and other 305 limitations). 306 307 Some architectures (e.g. ``x86``) may attempt to provide information 308 more precisely reflecting the current CPU frequency through this 309 attribute, but that still may not be the exact current CPU frequency as 310 seen by the hardware at the moment. This behavior though, is only 311 available via c:macro:``CPUFREQ_ARCH_CUR_FREQ`` option. 312 313``scaling_driver`` 314 The scaling driver currently in use. 315 316``scaling_governor`` 317 The scaling governor currently attached to this policy or (if the 318 |intel_pstate| scaling driver is in use) the scaling algorithm 319 provided by the driver that is currently applied to this policy. 320 321 This attribute is read-write and writing to it will cause a new scaling 322 governor to be attached to this policy or a new scaling algorithm 323 provided by the scaling driver to be applied to it (in the 324 |intel_pstate| case), as indicated by the string written to this 325 attribute (which must be one of the names listed by the 326 ``scaling_available_governors`` attribute described above). 327 328``scaling_max_freq`` 329 Maximum frequency the CPUs belonging to this policy are allowed to be 330 running at (in kHz). 331 332 This attribute is read-write and writing a string representing an 333 integer to it will cause a new limit to be set (it must not be lower 334 than the value of the ``scaling_min_freq`` attribute). 335 336``scaling_min_freq`` 337 Minimum frequency the CPUs belonging to this policy are allowed to be 338 running at (in kHz). 339 340 This attribute is read-write and writing a string representing a 341 non-negative integer to it will cause a new limit to be set (it must not 342 be higher than the value of the ``scaling_max_freq`` attribute). 343 344``scaling_setspeed`` 345 This attribute is functional only if the `userspace`_ scaling governor 346 is attached to the given policy. 347 348 It returns the last frequency requested by the governor (in kHz) or can 349 be written to in order to set a new frequency for the policy. 350 351 352Generic Scaling Governors 353========================= 354 355``CPUFreq`` provides generic scaling governors that can be used with all 356scaling drivers. As stated before, each of them implements a single, possibly 357parametrized, performance scaling algorithm. 358 359Scaling governors are attached to policy objects and different policy objects 360can be handled by different scaling governors at the same time (although that 361may lead to suboptimal results in some cases). 362 363The scaling governor for a given policy object can be changed at any time with 364the help of the ``scaling_governor`` policy attribute in ``sysfs``. 365 366Some governors expose ``sysfs`` attributes to control or fine-tune the scaling 367algorithms implemented by them. Those attributes, referred to as governor 368tunables, can be either global (system-wide) or per-policy, depending on the 369scaling driver in use. If the driver requires governor tunables to be 370per-policy, they are located in a subdirectory of each policy directory. 371Otherwise, they are located in a subdirectory under 372:file:`/sys/devices/system/cpu/cpufreq/`. In either case the name of the 373subdirectory containing the governor tunables is the name of the governor 374providing them. 375 376``performance`` 377--------------- 378 379When attached to a policy object, this governor causes the highest frequency, 380within the ``scaling_max_freq`` policy limit, to be requested for that policy. 381 382The request is made once at that time the governor for the policy is set to 383``performance`` and whenever the ``scaling_max_freq`` or ``scaling_min_freq`` 384policy limits change after that. 385 386``powersave`` 387------------- 388 389When attached to a policy object, this governor causes the lowest frequency, 390within the ``scaling_min_freq`` policy limit, to be requested for that policy. 391 392The request is made once at that time the governor for the policy is set to 393``powersave`` and whenever the ``scaling_max_freq`` or ``scaling_min_freq`` 394policy limits change after that. 395 396``userspace`` 397------------- 398 399This governor does not do anything by itself. Instead, it allows user space 400to set the CPU frequency for the policy it is attached to by writing to the 401``scaling_setspeed`` attribute of that policy. Though the intention may be to 402set an exact frequency for the policy, the actual frequency may vary depending 403on hardware coordination, thermal and power limits, and other factors. 404 405``schedutil`` 406------------- 407 408This governor uses CPU utilization data available from the CPU scheduler. It 409generally is regarded as a part of the CPU scheduler, so it can access the 410scheduler's internal data structures directly. 411 412It runs entirely in scheduler context, although in some cases it may need to 413invoke the scaling driver asynchronously when it decides that the CPU frequency 414should be changed for a given policy (that depends on whether or not the driver 415is capable of changing the CPU frequency from scheduler context). 416 417The actions of this governor for a particular CPU depend on the scheduling class 418invoking its utilization update callback for that CPU. If it is invoked by the 419RT or deadline scheduling classes, the governor will increase the frequency to 420the allowed maximum (that is, the ``scaling_max_freq`` policy limit). In turn, 421if it is invoked by the CFS scheduling class, the governor will use the 422Per-Entity Load Tracking (PELT) metric for the root control group of the 423given CPU as the CPU utilization estimate (see the *Per-entity load tracking* 424LWN.net article [1]_ for a description of the PELT mechanism). Then, the new 425CPU frequency to apply is computed in accordance with the formula 426 427 f = 1.25 * ``f_0`` * ``util`` / ``max`` 428 429where ``util`` is the PELT number, ``max`` is the theoretical maximum of 430``util``, and ``f_0`` is either the maximum possible CPU frequency for the given 431policy (if the PELT number is frequency-invariant), or the current CPU frequency 432(otherwise). 433 434This governor also employs a mechanism allowing it to temporarily bump up the 435CPU frequency for tasks that have been waiting on I/O most recently, called 436"IO-wait boosting". That happens when the :c:macro:`SCHED_CPUFREQ_IOWAIT` flag 437is passed by the scheduler to the governor callback which causes the frequency 438to go up to the allowed maximum immediately and then draw back to the value 439returned by the above formula over time. 440 441This governor exposes only one tunable: 442 443``rate_limit_us`` 444 Minimum time (in microseconds) that has to pass between two consecutive 445 runs of governor computations (default: 1.5 times the scaling driver's 446 transition latency or the maximum 2ms). 447 448 The purpose of this tunable is to reduce the scheduler context overhead 449 of the governor which might be excessive without it. 450 451This governor generally is regarded as a replacement for the older `ondemand`_ 452and `conservative`_ governors (described below), as it is simpler and more 453tightly integrated with the CPU scheduler, its overhead in terms of CPU context 454switches and similar is less significant, and it uses the scheduler's own CPU 455utilization metric, so in principle its decisions should not contradict the 456decisions made by the other parts of the scheduler. 457 458``ondemand`` 459------------ 460 461This governor uses CPU load as a CPU frequency selection metric. 462 463In order to estimate the current CPU load, it measures the time elapsed between 464consecutive invocations of its worker routine and computes the fraction of that 465time in which the given CPU was not idle. The ratio of the non-idle (active) 466time to the total CPU time is taken as an estimate of the load. 467 468If this governor is attached to a policy shared by multiple CPUs, the load is 469estimated for all of them and the greatest result is taken as the load estimate 470for the entire policy. 471 472The worker routine of this governor has to run in process context, so it is 473invoked asynchronously (via a workqueue) and CPU P-states are updated from 474there if necessary. As a result, the scheduler context overhead from this 475governor is minimum, but it causes additional CPU context switches to happen 476relatively often and the CPU P-state updates triggered by it can be relatively 477irregular. Also, it affects its own CPU load metric by running code that 478reduces the CPU idle time (even though the CPU idle time is only reduced very 479slightly by it). 480 481It generally selects CPU frequencies proportional to the estimated load, so that 482the value of the ``cpuinfo_max_freq`` policy attribute corresponds to the load of 4831 (or 100%), and the value of the ``cpuinfo_min_freq`` policy attribute 484corresponds to the load of 0, unless when the load exceeds a (configurable) 485speedup threshold, in which case it will go straight for the highest frequency 486it is allowed to use (the ``scaling_max_freq`` policy limit). 487 488This governor exposes the following tunables: 489 490``sampling_rate`` 491 This is how often the governor's worker routine should run, in 492 microseconds. 493 494 Typically, it is set to values of the order of 2000 (2 ms). Its 495 default value is to add a 50% breathing room 496 to ``cpuinfo_transition_latency`` on each policy this governor is 497 attached to. The minimum is typically the length of two scheduler 498 ticks. 499 500 If this tunable is per-policy, the following shell command sets the time 501 represented by it to be 1.5 times as high as the transition latency 502 (the default):: 503 504 # echo `$(($(cat cpuinfo_transition_latency) * 3 / 2))` > ondemand/sampling_rate 505 506``up_threshold`` 507 If the estimated CPU load is above this value (in percent), the governor 508 will set the frequency to the maximum value allowed for the policy. 509 Otherwise, the selected frequency will be proportional to the estimated 510 CPU load. 511 512``ignore_nice_load`` 513 If set to 1 (default 0), it will cause the CPU load estimation code to 514 treat the CPU time spent on executing tasks with "nice" levels greater 515 than 0 as CPU idle time. 516 517 This may be useful if there are tasks in the system that should not be 518 taken into account when deciding what frequency to run the CPUs at. 519 Then, to make that happen it is sufficient to increase the "nice" level 520 of those tasks above 0 and set this attribute to 1. 521 522``sampling_down_factor`` 523 Temporary multiplier, between 1 (default) and 100 inclusive, to apply to 524 the ``sampling_rate`` value if the CPU load goes above ``up_threshold``. 525 526 This causes the next execution of the governor's worker routine (after 527 setting the frequency to the allowed maximum) to be delayed, so the 528 frequency stays at the maximum level for a longer time. 529 530 Frequency fluctuations in some bursty workloads may be avoided this way 531 at the cost of additional energy spent on maintaining the maximum CPU 532 capacity. 533 534``powersave_bias`` 535 Reduction factor to apply to the original frequency target of the 536 governor (including the maximum value used when the ``up_threshold`` 537 value is exceeded by the estimated CPU load) or sensitivity threshold 538 for the AMD frequency sensitivity powersave bias driver 539 (:file:`drivers/cpufreq/amd_freq_sensitivity.c`), between 0 and 1000 540 inclusive. 541 542 If the AMD frequency sensitivity powersave bias driver is not loaded, 543 the effective frequency to apply is given by 544 545 f * (1 - ``powersave_bias`` / 1000) 546 547 where f is the governor's original frequency target. The default value 548 of this attribute is 0 in that case. 549 550 If the AMD frequency sensitivity powersave bias driver is loaded, the 551 value of this attribute is 400 by default and it is used in a different 552 way. 553 554 On Family 16h (and later) AMD processors there is a mechanism to get a 555 measured workload sensitivity, between 0 and 100% inclusive, from the 556 hardware. That value can be used to estimate how the performance of the 557 workload running on a CPU will change in response to frequency changes. 558 559 The performance of a workload with the sensitivity of 0 (memory-bound or 560 IO-bound) is not expected to increase at all as a result of increasing 561 the CPU frequency, whereas workloads with the sensitivity of 100% 562 (CPU-bound) are expected to perform much better if the CPU frequency is 563 increased. 564 565 If the workload sensitivity is less than the threshold represented by 566 the ``powersave_bias`` value, the sensitivity powersave bias driver 567 will cause the governor to select a frequency lower than its original 568 target, so as to avoid over-provisioning workloads that will not benefit 569 from running at higher CPU frequencies. 570 571``conservative`` 572---------------- 573 574This governor uses CPU load as a CPU frequency selection metric. 575 576It estimates the CPU load in the same way as the `ondemand`_ governor described 577above, but the CPU frequency selection algorithm implemented by it is different. 578 579Namely, it avoids changing the frequency significantly over short time intervals 580which may not be suitable for systems with limited power supply capacity (e.g. 581battery-powered). To achieve that, it changes the frequency in relatively 582small steps, one step at a time, up or down - depending on whether or not a 583(configurable) threshold has been exceeded by the estimated CPU load. 584 585This governor exposes the following tunables: 586 587``freq_step`` 588 Frequency step in percent of the maximum frequency the governor is 589 allowed to set (the ``scaling_max_freq`` policy limit), between 0 and 590 100 (5 by default). 591 592 This is how much the frequency is allowed to change in one go. Setting 593 it to 0 will cause the default frequency step (5 percent) to be used 594 and setting it to 100 effectively causes the governor to periodically 595 switch the frequency between the ``scaling_min_freq`` and 596 ``scaling_max_freq`` policy limits. 597 598``down_threshold`` 599 Threshold value (in percent, 20 by default) used to determine the 600 frequency change direction. 601 602 If the estimated CPU load is greater than this value, the frequency will 603 go up (by ``freq_step``). If the load is less than this value (and the 604 ``sampling_down_factor`` mechanism is not in effect), the frequency will 605 go down. Otherwise, the frequency will not be changed. 606 607``sampling_down_factor`` 608 Frequency decrease deferral factor, between 1 (default) and 10 609 inclusive. 610 611 It effectively causes the frequency to go down ``sampling_down_factor`` 612 times slower than it ramps up. 613 614 615Frequency Boost Support 616======================= 617 618Background 619---------- 620 621Some processors support a mechanism to raise the operating frequency of some 622cores in a multicore package temporarily (and above the sustainable frequency 623threshold for the whole package) under certain conditions, for example if the 624whole chip is not fully utilized and below its intended thermal or power budget. 625 626Different names are used by different vendors to refer to this functionality. 627For Intel processors it is referred to as "Turbo Boost", AMD calls it 628"Turbo-Core" or (in technical documentation) "Core Performance Boost" and so on. 629As a rule, it also is implemented differently by different vendors. The simple 630term "frequency boost" is used here for brevity to refer to all of those 631implementations. 632 633The frequency boost mechanism may be either hardware-based or software-based. 634If it is hardware-based (e.g. on x86), the decision to trigger the boosting is 635made by the hardware (although in general it requires the hardware to be put 636into a special state in which it can control the CPU frequency within certain 637limits). If it is software-based (e.g. on ARM), the scaling driver decides 638whether or not to trigger boosting and when to do that. 639 640The ``boost`` File in ``sysfs`` 641------------------------------- 642 643This file is located under :file:`/sys/devices/system/cpu/cpufreq/` and controls 644the "boost" setting for the whole system. It is not present if the underlying 645scaling driver does not support the frequency boost mechanism (or supports it, 646but provides a driver-specific interface for controlling it, like 647|intel_pstate|). 648 649If the value in this file is 1, the frequency boost mechanism is enabled. This 650means that either the hardware can be put into states in which it is able to 651trigger boosting (in the hardware-based case), or the software is allowed to 652trigger boosting (in the software-based case). It does not mean that boosting 653is actually in use at the moment on any CPUs in the system. It only means a 654permission to use the frequency boost mechanism (which still may never be used 655for other reasons). 656 657If the value in this file is 0, the frequency boost mechanism is disabled and 658cannot be used at all. 659 660The only values that can be written to this file are 0 and 1. 661 662Rationale for Boost Control Knob 663-------------------------------- 664 665The frequency boost mechanism is generally intended to help to achieve optimum 666CPU performance on time scales below software resolution (e.g. below the 667scheduler tick interval) and it is demonstrably suitable for many workloads, but 668it may lead to problems in certain situations. 669 670For this reason, many systems make it possible to disable the frequency boost 671mechanism in the platform firmware (BIOS) setup, but that requires the system to 672be restarted for the setting to be adjusted as desired, which may not be 673practical at least in some cases. For example: 674 675 1. Boosting means overclocking the processor, although under controlled 676 conditions. Generally, the processor's energy consumption increases 677 as a result of increasing its frequency and voltage, even temporarily. 678 That may not be desirable on systems that switch to power sources of 679 limited capacity, such as batteries, so the ability to disable the boost 680 mechanism while the system is running may help there (but that depends on 681 the workload too). 682 683 2. In some situations deterministic behavior is more important than 684 performance or energy consumption (or both) and the ability to disable 685 boosting while the system is running may be useful then. 686 687 3. To examine the impact of the frequency boost mechanism itself, it is useful 688 to be able to run tests with and without boosting, preferably without 689 restarting the system in the meantime. 690 691 4. Reproducible results are important when running benchmarks. Since 692 the boosting functionality depends on the load of the whole package, 693 single-thread performance may vary because of it which may lead to 694 unreproducible results sometimes. That can be avoided by disabling the 695 frequency boost mechanism before running benchmarks sensitive to that 696 issue. 697 698Legacy AMD ``cpb`` Knob 699----------------------- 700 701The AMD powernow-k8 scaling driver supports a ``sysfs`` knob very similar to 702the global ``boost`` one. It is used for disabling/enabling the "Core 703Performance Boost" feature of some AMD processors. 704 705If present, that knob is located in every ``CPUFreq`` policy directory in 706``sysfs`` (:file:`/sys/devices/system/cpu/cpufreq/policyX/`) and is called 707``cpb``, which indicates a more fine grained control interface. The actual 708implementation, however, works on the system-wide basis and setting that knob 709for one policy causes the same value of it to be set for all of the other 710policies at the same time. 711 712That knob is still supported on AMD processors that support its underlying 713hardware feature, but it may be configured out of the kernel (via the 714:c:macro:`CONFIG_X86_ACPI_CPUFREQ_CPB` configuration option) and the global 715``boost`` knob is present regardless. Thus it is always possible use the 716``boost`` knob instead of the ``cpb`` one which is highly recommended, as that 717is more consistent with what all of the other systems do (and the ``cpb`` knob 718may not be supported any more in the future). 719 720The ``cpb`` knob is never present for any processors without the underlying 721hardware feature (e.g. all Intel ones), even if the 722:c:macro:`CONFIG_X86_ACPI_CPUFREQ_CPB` configuration option is set. 723 724 725References 726========== 727 728.. [1] Jonathan Corbet, *Per-entity load tracking*, 729 https://lwn.net/Articles/531853/ 730