我正在努力解决 intel_pstate 性能调节器未以最大 CPU 速度运行的问题。我有一个 2.2ghz 系统，出于正当理由它需要全速运行。
# grep -E '^model name|^cpu MHz' /proc/cpuinfo
model name : Intel(R) Xeon(R) CPU E5-2618L v4 @ 2.20GHz
cpu MHz : 1393.933
model name : Intel(R) Xeon(R) CPU E5-2618L v4 @ 2.20GHz
cpu MHz : 1344.519
model name : Intel(R) Xeon(R) CPU E5-2618L v4 @ 2.20GHz
cpu MHz : 1379.833
model name : Intel(R) Xeon(R) CPU E5-2618L v4 @ 2.20GHz
cpu MHz : 1398.901
model name : Intel(R) Xeon(R) CPU E5-2618L v4 @ 2.20GHz
cpu MHz : 1352.709
model name : Intel(R) Xeon(R) CPU E5-2618L v4 @ 2.20GHz
cpu MHz : 1395.007
model name : Intel(R) Xeon(R) CPU E5-2618L v4 @ 2.20GHz
cpu MHz : 1356.738
model name : Intel(R) Xeon(R) CPU E5-2618L v4 @ 2.20GHz
cpu MHz : 1350.292
model name : Intel(R) Xeon(R) CPU E5-2618L v4 @ 2.20GHz
cpu MHz : 1349.890
model name : Intel(R) Xeon(R) CPU E5-2618L v4 @ 2.20GHz
cpu MHz : 1354.589
model name : Intel(R) Xeon(R) CPU E5-2618L v4 @ 2.20GHz
cpu MHz : 1554.260
model name : Intel(R) Xeon(R) CPU E5-2618L v4 @ 2.20GHz
cpu MHz : 1513.439
model name : Intel(R) Xeon(R) CPU E5-2618L v4 @ 2.20GHz
cpu MHz : 1500.952
model name : Intel(R) Xeon(R) CPU E5-2618L v4 @ 2.20GHz
cpu MHz : 1466.442
model name : Intel(R) Xeon(R) CPU E5-2618L v4 @ 2.20GHz
cpu MHz : 1515.856
model name : Intel(R) Xeon(R) CPU E5-2618L v4 @ 2.20GHz
cpu MHz : 1552.917
model name : Intel(R) Xeon(R) CPU E5-2618L v4 @ 2.20GHz
cpu MHz : 1495.178
model name : Intel(R) Xeon(R) CPU E5-2618L v4 @ 2.20GHz
cpu MHz : 1476.245
model name : Intel(R) Xeon(R) CPU E5-2618L v4 @ 2.20GHz
cpu MHz : 1518.945
model name : Intel(R) Xeon(R) CPU E5-2618L v4 @ 2.20GHz
cpu MHz : 1493.566
cpupower 显示表现州长，但频率没有调整：
# cpupower frequency-info
analyzing CPU 0:
driver: intel_pstate
CPUs which run at the same hardware frequency: 0
CPUs which need to have their frequency coordinated by software: 0
maximum transition latency: Cannot determine or is not supported.
hardware limits: 1.20 GHz - 2.20 GHz
available cpufreq governors: performance powersave
current policy: frequency should be within 2.00 GHz and 2.20 GHz.
The governor "performance" may decide which speed to use
within this range.
current CPU frequency: 1.35 GHz (asserted by call to hardware)
boost state support:
Supported: no
Active: no
tuned-adm 列表显示所选的吞吐量性能调控器：
# tuned-adm list
Available profiles:
- balanced - General non-specialized tuned profile
- desktop - Optimize for the desktop use-case
- hpc-compute - Optimize for HPC compute workloads
- latency-performance - Optimize for deterministic performance at the cost of increased power consumption
- network-latency - Optimize for deterministic performance at the cost of increased power consumption, focused on low latency network performance
- network-throughput - Optimize for streaming network throughput, generally only necessary on older CPUs or 40G+ networks
- powersave - Optimize for low power consumption
- throughput-performance - Broadly applicable tuning that provides excellent performance across a variety of common server workloads
- virtual-guest - Optimize for running inside a virtual guest
- virtual-host - Optimize for running KVM guests
Current active profile: throughput-performance
cpufreq 文件系统显示最小值未调整且不支持“setspeed”：
grep . /sys/devices/system/cpu/cpu0/cpufreq/*
/sys/devices/system/cpu/cpu0/cpufreq/affected_cpus:0
/sys/devices/system/cpu/cpu0/cpufreq/cpuinfo_cur_freq:1439721
/sys/devices/system/cpu/cpu0/cpufreq/cpuinfo_max_freq:2200000
/sys/devices/system/cpu/cpu0/cpufreq/cpuinfo_min_freq:1200000
/sys/devices/system/cpu/cpu0/cpufreq/cpuinfo_transition_latency:4294967295
/sys/devices/system/cpu/cpu0/cpufreq/related_cpus:0
/sys/devices/system/cpu/cpu0/cpufreq/scaling_available_governors:performance powersave
/sys/devices/system/cpu/cpu0/cpufreq/scaling_cur_freq:1439721
/sys/devices/system/cpu/cpu0/cpufreq/scaling_driver:intel_pstate
/sys/devices/system/cpu/cpu0/cpufreq/scaling_governor:performance
/sys/devices/system/cpu/cpu0/cpufreq/scaling_max_freq:2200000
/sys/devices/system/cpu/cpu0/cpufreq/scaling_min_freq:2000000
/sys/devices/system/cpu/cpu0/cpufreq/scaling_setspeed:<unsupported>
我已经能够禁用 intel_pstate，这会导致重新启动时 CPU 频率达到最大，但我希望将系统调低至 1.9 GHz 进行测试。当我禁用 intel_pstate 时，acpi 缩放不可用，并且 cpupower 无法执行任何操作。
相关信息：
- 操作系统版本：RHEL 7.7
- 内核：3.10.0-1062.9.1.rt56.1033.el7.x86_64 #1 SMP PREEMPT RT 2019 年 12 月 2 日周一 12:20:53 EST 2019 x86_64 x86_64 x86_64 GNU/Linux
- cpufreq_ondemand：不可用
- 热节流阀：在各个 cpu* 文件夹中全部设置为零
- 邮件信息：
# dmesg -t --level=alert,crit,err,warn
ACPI: RSDP 00000000000f05b0 00024 (v02 ALASKA)
ACPI: XSDT 00000000799820a0 000C4 (v01 ALASKA A M I 01072009 AMI 00010013)
ACPI: FACP 00000000799b4ed8 0010C (v05 ALASKA A M I 01072009 AMI 00010013)
ACPI: DSDT 0000000079982200 32CD7 (v02 ALASKA A M I 01072009 INTL 20091013)
ACPI: FACS 0000000079e5bf80 00040
ACPI: APIC 00000000799b4fe8 0017C (v03 ALASKA A M I 01072009 AMI 00010013)
ACPI: FPDT 00000000799b5168 00044 (v01 ALASKA A M I 01072009 AMI 00010013)
ACPI: FIDT 00000000799b51b0 0009C (v01 ALASKA A M I 01072009 AMI 00010013)
ACPI: TCPA 00000000799b5250 00032 (v02 ALASKA NAPAASF 00000000 MSFT 01000013)
ACPI: MCFG 00000000799b5288 0003C (v01 ALASKA A M I 01072009 MSFT 00000097)
ACPI: UEFI 00000000799b52c8 00042 (v01 ALASKA A M I 01072009 00000000)
ACPI: HPET 00000000799b5310 00038 (v01 ALASKA A M I 00000001 INTL 20091013)
ACPI: MSCT 00000000799b5348 00090 (v01 ALASKA A M I 00000001 INTL 20091013)
ACPI: SLIT 00000000799b53d8 00030 (v01 ALASKA A M I 00000001 INTL 20091013)
ACPI: SRAT 00000000799b5408 01158 (v03 ALASKA A M I 00000001 INTL 20091013)
ACPI: WDDT 00000000799b6560 00040 (v01 ALASKA A M I 00000000 INTL 20091013)
ACPI: SSDT 00000000799b65a0 16FB3 (v02 ALASKA PmMgt 00000001 INTL 20120913)
ACPI: NITR 00000000799cd558 00071 (v02 ALASKA A M I 00000001 INTL 20091013)
ACPI: SPCR 00000000799cd5d0 00050 (v01 A M I APTIO V 01072009 AMI. 0005000B)
ACPI: TPM2 00000000799cd620 00034 (v03 Tpm2Tabl 00000001 AMI 00000000)
ACPI: SSDT 00000000799cd658 02652 (v02 ALASKA SpsNm 00000002 INTL 20120913)
ACPI: SSDT 00000000799cfcb0 00064 (v02 ALASKA SpsNvs 00000002 INTL 20120913)
ACPI: PRAD 00000000799cfd18 00102 (v02 ALASKA A M I 00000002 INTL 20120913)
ACPI: DMAR 00000000799cfe20 00108 (v01 ALASKA A M I 00000001 INTL 20091013)
Zone ranges:
DMA [mem 0x00001000-0x00ffffff]
DMA32 [mem 0x01000000-0xffffffff]
Normal [mem 0x100000000-0x207fffffff]
Movable zone start for each node
Early memory node ranges
node 0: [mem 0x00001000-0x00098fff]
node 0: [mem 0x00100000-0x7908efff]
node 0: [mem 0x100000000-0x107fffffff]
node 1: [mem 0x1080000000-0x207fffffff]
Built 2 zonelists in Zone order, mobility grouping on. Total pages: 33001935
Policy zone: Normal
ENERGY_PERF_BIAS: Set to 'normal', was 'performance'
ENERGY_PERF_BIAS: View and update with x86_energy_perf_policy(8)
ACPI: All ACPI Tables successfully acquired
ACPI: Enabled 5 GPEs in block 00 to 3F
KSM thread running on realtime kernel. This may introduce latency spikes...
Dquot-cache hash table entries: 512 (order 0, 4096 bytes)
i8042: No controller found
systemd[1]: [/run/systemd/generator/dev-mapper-vg00\x2dlv_root.device.d/timeout.conf:3] Unknown lvalue 'JobRunningTimeoutSec' in section 'Unit'
mlx5_port_module_event: 3 callbacks suppressed
tsyncpci: loading out-of-tree module taints kernel.
Spectracom TSync Timing Board - version 3.12
Copyright (c) 2013 Spectracom Corporation
TSYNC version of board: tsyncpci0
audit_printk_skb: 156 callbacks suppressed
ibm4765: ibm4765, version 2.1.1, build 8 built at Jan 28 2020 18:24:36 initializing
intel_pstate: Turbo disabled by BIOS or unavailable on processor
GRUB 默认值：
GRUB_CMDLINE_LINUX_DEFAULT="crashkernel=auto rd.shell=0 rd.auto=1 audit=1 acpi_pad.disable=1 nmi_watchdog=1 boot=/dev/sda2 console=ttyS0,115200n8 console=ttyS1 text"