baseline power consumtion kernel > 5.3.10

baseline power consumtion kernel > 5.3.10

[Guido Riedel]

Hi,

Since Kernel 5.3.11 I got a significant increased baseline Power 
consumption. I tried up to 5.4.1 on arch linux. Powertop shows an 
increase from 3.6W on 5.3.10 to 5.x W on higher Kernels. My laptop fan 
spins a lot more.

I read that a lot of Intel stuff has entered the Kernel in 5.3.11, could 
this cause it? Will it stay like this in the next Versions? Can I maybe 
switch something off etc. to get a lower baseline power?

I hope I write to the right address, otherwise sorry.

Greetings Guido

lscpu:

Architektur:                     x86_64
CPU Operationsmodus:             32-bit, 64-bit
Byte-Reihenfolge:                Little Endian
Adressgrößen:                    39 bits physical, 48 bits virtual
CPU(s):                          8
Liste der Online-CPU(s):         0-7
Thread(s) pro Kern:              2
Kern(e) pro Socket:              4
Sockel:                          1
NUMA-Knoten:                     1
Anbieterkennung:                 GenuineIntel
Prozessorfamilie:                6
Modell:                          142
Modellname:                      Intel(R) Core(TM) i7-8550U CPU @ 1.80GHz
Stepping:                        10
CPU MHz:                         1072.191
Maximale Taktfrequenz der CPU:   4000,0000
Minimale Taktfrequenz der CPU:   400,0000
BogoMIPS:                        4001.60
Virtualisierung:                 VT-x
L1d Cache:                       128 KiB
L1i Cache:                       128 KiB
L2 Cache:                        1 MiB
L3 Cache:                        8 MiB
NUMA-Knoten0 CPU(s):             0-7
Vulnerability L1tf:              Mitigation; PTE Inversion; VMX 
conditional cache flushes, SMT vulnerable
Vulnerability Mds:               Vulnerable: Clear CPU buffers 
attempted, no microcode; SMT vulnerable
Vulnerability Meltdown:          Mitigation; PTI
Vulnerability Spec store bypass: Mitigation; Speculative Store Bypass 
disabled via prctl and seccomp
Vulnerability Spectre v1:        Mitigation; usercopy/swapgs barriers 
and __user pointer sanitization
Vulnerability Spectre v2:        Mitigation; Full generic retpoline, 
IBPB conditional, IBRS_FW, STIBP conditional, RSB filling
Markierungen:                    fpu vme de pse tsc msr pae mce cx8 apic 
sep mtrr pge mca cmov pat pse36 clflush dts acpi mmx fxsr s
                                  se sse2 ss ht tm pbe syscall nx 
pdpe1gb rdtscp lm constant_tsc art arch_perfmon pebs bts rep_good n
                                  opl xtopology nonstop_tsc cpuid 
aperfmperf pni pclmulqdq dtes64 monitor ds_cpl vmx est tm2 ssse3 sd
                                  bg fma cx16 xtpr pdcm pcid sse4_1 
sse4_2 x2apic movbe popcnt tsc_deadline_timer aes xsave avx f16c
                                  rdrand lahf_lm abm 3dnowprefetch 
cpuid_fault epb invpcid_single pti ssbd ibrs ibpb stibp tpr_shadow
                                   vnmi flexpriority ept vpid ept_ad 
fsgsbase tsc_adjust bmi1 avx2 smep bmi2 erms invpcid mpx rdseed
                                  adx smap clflushopt intel_pt xsaveopt 
xsavec xgetbv1 xsaves dtherm ida arat pln pts hwp hwp_notify
                                  hwp_act_window hwp_epp flush_l1d

Re: baseline power consumtion kernel > 5.3.10

[Rafael J. Wysocki]

On Thu, Dec 5, 2019 at 8:52 AM Guido Riedel <guido.riedel@web.de> wrote:
>
> Hi,
>
> Since Kernel 5.3.11 I got a significant increased baseline Power
> consumption. I tried up to 5.4.1 on arch linux. Powertop shows an
> increase from 3.6W on 5.3.10 to 5.x W on higher Kernels. My laptop fan
> spins a lot more.
>
> I read that a lot of Intel stuff has entered the Kernel in 5.3.11, could
> this cause it? Will it stay like this in the next Versions? Can I maybe
> switch something off etc. to get a lower baseline power?

That's possible.

It would be good to compare the general activity of the system
depending on which version of the kernel is in use.

The turbostat utility can be used for that.

It is included in the kernel source tree (tools/power/x86/turbostat/).
Build it, install it (the version from 5.4.1 can be used with older
kernels too), run it along with your normal workload for some time
with both 5.3.11 and 5.4.1, like this:

# turbostat --quiet --out=output_file sleep 600

(this will cause it to run until the "sleep 600" command is complete,
i.e. for 10 minutes).  Then post the output files from both runs.

Note that you need to run it as root (there is a man page etc).

Thanks!

Re: baseline power consumtion kernel > 5.3.10

[Lucas Stach]

On Do, 2019-12-05 at 08:52 +0100, Guido Riedel wrote:
> Hi,
> 
> Since Kernel 5.3.11 I got a significant increased baseline Power 
> consumption. I tried up to 5.4.1 on arch linux. Powertop shows an 
> increase from 3.6W on 5.3.10 to 5.x W on higher Kernels. My laptop fan 
> spins a lot more.
> 
> I read that a lot of Intel stuff has entered the Kernel in 5.3.11, could 
> this cause it? Will it stay like this in the next Versions? Can I maybe 
> switch something off etc. to get a lower baseline power?
> 
> I hope I write to the right address, otherwise sorry.

7e34f4e4aad3 ("drm/i915/gen8+: Add RC6 CTX corruption WA"), which has
been backported to 5.3.11 causes a major power regression due to Intel
GPUs not being able to enter rc6 anymore.

Fixes for this are -next, but I don't know what the plans for landing
and maybe backporting are.

Regards,
Lucas

Re: baseline power consumtion kernel > 5.3.10

[Lucas Stach]

On Do, 2019-12-05 at 08:52 +0100, Guido Riedel wrote:
> Hi,
> 
> Since Kernel 5.3.11 I got a significant increased baseline Power 
> consumption. I tried up to 5.4.1 on arch linux. Powertop shows an 
> increase from 3.6W on 5.3.10 to 5.x W on higher Kernels. My laptop fan 
> spins a lot more.
> 
> I read that a lot of Intel stuff has entered the Kernel in 5.3.11, could 
> this cause it? Will it stay like this in the next Versions? Can I maybe 
> switch something off etc. to get a lower baseline power?
> 
> I hope I write to the right address, otherwise sorry.

7e34f4e4aad3 ("drm/i915/gen8+: Add RC6 CTX corruption WA"), which has
been backported to 5.3.11 causes a major power regression due to Intel
GPUs not being able to enter rc6 anymore.

Fixes for this are -next, but I don't know what the plans for landing
and maybe backporting are.

Regards,
Lucas

_______________________________________________
dri-devel mailing list
dri-devel@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/dri-devel