| /linux-5.10/drivers/acpi/ |
| D | processor_perflib.c | 26 #define ACPI_PROCESSOR_FILE_PERFORMANCE "performance"
83 if (ppc >= pr->performance->state_count || in acpi_processor_get_platform_limit()
88 pr->performance->states[ppc].core_frequency * 1000); in acpi_processor_get_platform_limit()
102 * 0: success. OSPM is now using the performance state specificed.
116 if (ignore_ppc || !pr->performance) { in acpi_processor_ppc_has_changed()
146 if (!pr || !pr->performance || !pr->performance->state_count) in acpi_processor_get_bios_limit()
148 *limit = pr->performance->states[pr->performance_platform_limit]. in acpi_processor_get_bios_limit()
228 memcpy(&pr->performance->control_register, obj.buffer.pointer, in acpi_processor_get_performance_control()
245 memcpy(&pr->performance->status_register, obj.buffer.pointer, in acpi_processor_get_performance_control()
314 ACPI_DEBUG_PRINT((ACPI_DB_INFO, "Found %d performance states\n", in acpi_processor_get_performance_states()
[all …]
|
| /linux-5.10/Documentation/admin-guide/acpi/ |
| D | cppc_sysfs.rst | 4 Collaborative Processor Performance Control (CPPC)
11 performance of a logical processor on a contigious and abstract performance
12 scale. CPPC exposes a set of registers to describe abstract performance scale,
13 to request performance levels and to measure per-cpu delivered performance.
38 * highest_perf : Highest performance of this processor (abstract scale).
39 * nominal_perf : Highest sustained performance of this processor
41 * lowest_nonlinear_perf : Lowest performance of this processor with nonlinear
43 * lowest_perf : Lowest performance of this processor (abstract scale).
47 The above frequencies should only be used to report processor performance in
51 * feedback_ctrs : Includes both Reference and delivered performance counter.
[all …]
|
| /linux-5.10/include/linux/ |
| D | energy_model.h | 14 * em_perf_state - Performance state of a performance domain
29 * em_perf_domain - Performance domain
30 * @table: List of performance states, in ascending order
31 * @nr_perf_states: Number of performance states
33 * for performance reasons to avoid potential cache
37 * In case of CPU device, a "performance domain" represents a group of CPUs
38 * whose performance is scaled together. All CPUs of a performance domain
39 * must have the same micro-architecture. Performance domains often have
56 * active_power() - Provide power at the next performance state of
58 * @power : Active power at the performance state in mW
[all …]
|
| /linux-5.10/tools/power/x86/x86_energy_perf_policy/ |
| D | x86_energy_perf_policy.8 | 5 x86_energy_perf_policy \- Manage Energy vs. Performance Policy via x86 Model Specific Registers
18 .RB "value: # | default | performance | balance-performance | balance-power | power"
21 displays and updates energy-performance policy settings specific to
28 and Processor Performance States (P-states).
31 Further, it allows the OS to influence energy/performance trade-offs where there
82 Set a policy with a normal balance between performance and energy efficiency.
83 The processor will tolerate minor performance compromise
88 .I performance
89 Set a policy for maximum performance,
90 accepting no performance sacrifice for the benefit of energy efficiency.
[all …]
|
| /linux-5.10/Documentation/admin-guide/pm/ |
| D | intel-speed-select.rst | 8 collection of features that give more granular control over CPU performance.
9 With Intel(R) SST, one server can be configured for power and performance for a
15 …tel.com/docs/networkbuilders/intel-speed-select-technology-base-frequency-enhancing-performance.pdf
25 how these commands change the power and performance profile of the system under
79 Intel(R) Speed Select Technology - Performance Profile (Intel(R) SST-PP)
83 performance requirements. This helps users during deployment as they do not have
85 Technology - Performance Profile (Intel(R) SST-PP) feature introduces a mechanism
86 that allows multiple optimized performance profiles per system. Each profile
89 performance profile and meet CPU online/offline requirement, the user can expect
93 Number or performance levels
[all …]
|
| D | intel_pstate.rst | 5 ``intel_pstate`` CPU Performance Scaling Driver
17 :doc:`CPU performance scaling subsystem <cpufreq>` in the Linux kernel
25 than just an operating frequency or an operating performance point (see the
30 uses frequencies for identifying operating performance points of CPUs and
58 active mode, it uses its own internal performance scaling governor algorithm or
61 a certain performance scaling algorithm. Which of them will be in effect
88 active mode: ``powersave`` and ``performance``. The way they both operate
94 Namely, if that option is set, the ``performance`` algorithm will be used by
117 HWP + ``performance``
121 Energy-Performance Preference (EPP) knob (if supported) or its
[all …]
|
| D | intel_epb.rst | 5 Intel Performance and Energy Bias Hint
16 Intel Performance and Energy Bias Attribute in ``sysfs``
19 The Intel Performance and Energy Bias Hint (EPB) value for a given (logical) CPU
26 a value of 0 corresponds to a hint preference for highest performance
31 with one of the strings: "performance", "balance-performance", "normal",
|
| D | cpufreq.rst | 7 CPU Performance Scaling
15 The Concept of CPU Performance Scaling
20 Operating Performance Points or P-states (in ACPI terminology). As a rule,
30 highest one (i.e. the highest-performance frequency/voltage configuration
44 to as CPU performance scaling or CPU frequency scaling (because it involves
48 CPU Performance Scaling in Linux
51 The Linux kernel supports CPU performance scaling by means of the ``CPUFreq``
56 interfaces for all platforms that support CPU performance scaling. It defines
70 performance scaling algorithms for P-state selection can be represented in a
72 to use the same performance scaling algorithm implemented in exactly the same
[all …]
|
| /linux-5.10/arch/powerpc/include/asm/ |
| D | reg_fsl_emb.h | 3 * Contains register definitions for the Freescale Embedded Performance
13 /* Performance Monitor Registers */
20 /* Freescale Book E Performance Monitor APU Registers */
21 #define PMRN_PMC0 0x010 /* Performance Monitor Counter 0 */
22 #define PMRN_PMC1 0x011 /* Performance Monitor Counter 1 */
23 #define PMRN_PMC2 0x012 /* Performance Monitor Counter 2 */
24 #define PMRN_PMC3 0x013 /* Performance Monitor Counter 3 */
25 #define PMRN_PMC4 0x014 /* Performance Monitor Counter 4 */
26 #define PMRN_PMC5 0x015 /* Performance Monitor Counter 5 */
67 #define PMRN_UPMC0 0x000 /* User Performance Monitor Counter 0 */
[all …]
|
| /linux-5.10/Documentation/admin-guide/ |
| D | perf-security.rst | 9 Usage of Performance Counters for Linux (perf_events) [1]_ , [2]_ , [3]_
14 depends on the nature of data that perf_events performance monitoring
15 units (PMU) [2]_ and Perf collect and expose for performance analysis.
16 Collected system and performance data may be split into several
21 its topology, used kernel and Perf versions, performance monitoring
30 faults, CPU migrations), architectural hardware performance counters
46 So, perf_events performance monitoring and observability operations are
56 all kernel security permission checks so perf_events performance
70 as privileged processes with respect to perf_events performance
73 privilege [13]_ (POSIX 1003.1e: 2.2.2.39) for performance monitoring and
[all …]
|
| /linux-5.10/drivers/gpu/drm/msm/disp/dpu1/ |
| D | dpu_core_perf.h | 31 * struct dpu_core_perf_params - definition of performance parameters
43 * struct dpu_core_perf_tune - definition of performance tuning control
44 * @mode: performance mode
55 * struct dpu_core_perf - definition of core performance context
62 * @perf_tune: debug control for performance tuning
83 * dpu_core_perf_crtc_check - validate performance of the given crtc state
92 * dpu_core_perf_crtc_update - update performance of the given crtc
108 * dpu_core_perf_destroy - destroy the given core performance context
109 * @perf: Pointer to core performance context
114 * dpu_core_perf_init - initialize the given core performance context
[all …]
|
| /linux-5.10/Documentation/power/ |
| D | energy-model.rst | 11 the power consumed by devices at various performance levels, and the kernel
53 'performance domain' in the system. A performance domain is a group of CPUs
54 whose performance is scaled together. Performance domains generally have a
55 1-to-1 mapping with CPUFreq policies. All CPUs in a performance domain are
56 required to have the same micro-architecture. CPUs in different performance
69 2.2 Registration of performance domains
72 Drivers are expected to register performance domains into the EM framework by
79 for each performance state. The callback function provided by the driver is free
82 performance domains using cpumask. For other devices than CPUs the last
89 2.3 Accessing performance domains
[all …]
|
| /linux-5.10/arch/x86/events/ |
| D | Kconfig | 2 menu "Performance monitoring"
5 tristate "Intel uncore performance events"
9 Include support for Intel uncore performance events. These are
13 tristate "Intel/AMD rapl performance events"
17 Include support for Intel and AMD rapl performance events for power
21 tristate "Intel cstate performance events"
25 Include support for Intel cstate performance events for power
|
| /linux-5.10/drivers/perf/ |
| D | Kconfig | 3 # Performance Monitor Drivers
6 menu "Performance monitor support"
56 Say y if you want to use CPU performance monitors on ARM-based
64 tristate "ARM SMMUv3 Performance Monitors Extension"
67 Provides support for the ARM SMMUv3 Performance Monitor Counter
76 Provides support for performance monitor unit in ARM DynamIQ Shared
85 Provides support for the DDR performance monitor in i.MX8, which
94 Provides support for the L2 cache performance monitor unit (PMU)
104 Provides support for the L3 cache performance monitor unit (PMU)
123 Say y if you want to use APM X-Gene SoC performance monitors.
|
| /linux-5.10/kernel/power/ |
| D | energy_model.c | 20 * Mutex serializing the registrations of performance domains and letting
60 /* Create the directory of the performance domain */ in em_debug_create_pd()
67 /* Create a sub-directory for each performance state */ in em_debug_create_pd()
107 /* Build the list of performance states for this performance domain */ in em_create_perf_table()
111 * lowest performance state of 'dev' above 'freq' and updates in em_create_perf_table()
123 * higher performance states. in em_create_perf_table()
157 /* Compute the cost of each performance state. */ in em_create_perf_table()
211 * em_pd_get() - Return the performance domain for a device
212 * @dev : Device to find the performance domain for
214 * Returns the performance domain to which @dev belongs, or NULL if it doesn't
[all …]
|
| /linux-5.10/tools/power/cpupower/bench/ |
| D | README-BENCH | 7 - Identify worst case performance loss when doing dynamic frequency
12 - Identify cpufreq related performance regressions between kernels
18 - Power saving related regressions (In fact as better the performance
28 For that purpose, it compares the performance governor to a configured
56 takes on this machine and needs to be run in a loop using the performance
58 Then the above test runs are processed using the performance governor
61 on full performance and you get the overall performance loss.
80 trigger of the cpufreq-bench, you will see no performance loss (compare with
84 will always see 50% loads and you get worst performance impact never
|
| D | benchmark.c | 70 * generates a specific sleep an load time with the performance
93 total_time *= 2; /* powersave and performance cycles */ in start_benchmark()
101 /* set the cpufreq governor to "performance" which disables in start_benchmark()
103 if (set_cpufreq_governor("performance", config->cpu) != 0) in start_benchmark()
124 /* do some sleep/load cycles with the performance governor */ in start_benchmark()
132 printf("performance cycle took %lius, " in start_benchmark()
175 printf("performance is at %.2f%%\n", in start_benchmark()
|
| /linux-5.10/Documentation/scheduler/ |
| D | sched-energy.rst | 38 performance [inst/s]
48 while still getting 'good' performance. It is essentially an alternative
49 optimization objective to the current performance-only objective for the
51 performance.
78 task/CPU is, and to take this into consideration when evaluating performance vs
84 per 'performance domain' in the system (see Documentation/power/energy-model.rst
85 for futher details about performance domains).
89 scheduler maintains a singly linked list of all performance domains intersecting
95 necessarily match those of performance domains, the lists of different root
99 Let us consider a platform with 12 CPUs, split in 3 performance domains
[all …]
|
| /linux-5.10/kernel/rcu/ |
| D | Kconfig.debug | 27 tristate "performance tests for RCU"
36 This option provides a kernel module that runs performance
40 Say Y here if you want RCU performance tests to be built into
42 Say M if you want the RCU performance tests to build as a module.
74 This option provides a kernel module that runs performance tests
79 Say Y here if you want these performance tests built into the kernel.
126 lifetime and kills performance. Don't try this on large
|
| /linux-5.10/tools/testing/selftests/vm/ |
| D | test_vmalloc.sh | 9 # a) analyse performance of vmalloc allocations;
22 # Static templates for performance, stressing and smoke tests.
51 echo "Run performance tests to evaluate how fast vmalloc allocation is."
74 echo "for deep performance analysis as well as stress testing."
83 echo -n "Usage: $0 [ performance ] | [ stress ] | | [ smoke ] | "
107 echo "# Performance analysis"
108 echo "./${DRIVER}.sh performance"
161 if [[ "$1" = "performance" ]]; then
|
| /linux-5.10/tools/power/cpupower/man/ |
| D | cpupower-set.1 | 27 its policy for the relative importance of performance versus energy savings to
31 performance and 15 is maximum energy efficiency.
34 when it must select trade-offs between performance and
37 This policy hint does not supersede Processor Performance states
51 Setting the performance bias value on one CPU can modify the setting on
|
| /linux-5.10/arch/x86/kernel/cpu/ |
| D | intel_epb.c | 3 * Intel Performance and Energy Bias Hint support.
25 * The Performance and Energy Bias Hint (EPB) allows software to specify its
26 * preference with respect to the power-performance tradeoffs present in the
47 * firmware is 0 ('performance') and at least on some of them that is because
87 * 0 ('performance') at this point, assume that it has not been in intel_epb_restore()
94 pr_warn_once("ENERGY_PERF_BIAS: Set to 'normal', was 'performance'\n"); in intel_epb_restore()
106 "performance",
107 "balance-performance",
|
| /linux-5.10/Documentation/admin-guide/mm/ |
| D | numaperf.rst | 9 as CPU cache coherence, but may have different performance. For example,
13 under different domains, or "nodes", based on locality and performance
35 performance when accessing a given memory target. Each initiator-target
55 nodes' access characteristics share the same performance relative to other
65 NUMA Performance
69 be allocated from based on the node's performance characteristics. If
79 The performance characteristics the kernel provides for the local initiators
104 performance characteristics in order to provide large address space of
130 attributes in order to maximize the performance out of such a setup.
|
| /linux-5.10/Documentation/networking/device_drivers/ethernet/neterion/ |
| D | s2io.rst | 14 - 5. Performance suggestions
63 significant performance improvement on certain platforms(SGI Altix,
67 (IA64, Xeon) resulting in noticeable performance improvement(up to 7%
114 5. Performance suggestions
123 good performance::
133 Transmit performance:
164 Receive performance:
173 b. Use 2-buffer mode. This results in large performance boost on
|
| /linux-5.10/drivers/perf/hisilicon/ |
| D | Kconfig | 6 Support for HiSilicon SoC L3 Cache performance monitor, Hydra Home
7 Agent performance monitor and DDR Controller performance monitor.
|