Hostfrequencyscaling #140
Merged
Hostfrequencyscaling #140
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
kpelzel
suggested changes
Dec 12, 2019
kraken/main.go.tpl
Outdated
Comment on lines
200
to
206
| k.Sme.Thaw() | ||
| // Thaw if full state | ||
| if len(parents) == 0 { | ||
| k.Sme.Thaw() | ||
| } |
There was a problem hiding this comment.
We might want this in a separate PR, but ask Lowell.
Comment on lines
66
to
76
| // PxeURL refers to PXE object | ||
| PxeURL string = "type.googleapis.com/proto.RPi3/Pxe" | ||
|
|
||
| // ModuleStateURL refers to module state | ||
| ModuleStateURL string = "/Services/hostfrequencyscaling/State" | ||
|
|
||
| // HostThermalStateURL points to Thermal extension | ||
| HostThermalStateURL string = "type.googleapis.com/proto.HostThermal/State" | ||
|
|
||
| // NodeIPURL provides node IP address | ||
| NodeIPURL string = "type.googleapis.com/proto.IPv4OverEthernet/Ifaces/0/Ip/Ip" |
There was a problem hiding this comment.
These constant values probably don't need to be exported, but it probably doesn't really matter. Ask Lowell
|
I just realized a lot of this PR contains changes from #112. We should make sure that gets merged first, then do a rebase. |
ghazanfarttu
force-pushed
the
hostfrequencyscaling
branch
5 times, most recently
from
7e0076f to
83f6e48
Compare
kpelzel
force-pushed
the
hostfrequencyscaling
branch
from
93f15ed to
a95b77b
Compare
kpelzel
suggested changes
Jan 7, 2020
config/pipxe-child.yaml
Outdated
Comment on lines
1
to
28
| # kraken-build.go: describes a build for a BitScope Raspberry Pi cluster | ||
| targets: | ||
| 'linux-arm64': # this identifies the build, will be appended to the binary name | ||
| os: 'linux' # os must match a supported GOOS | ||
| arch: 'arm64' # arch must match a supported GOARCH | ||
| 'linux-amd64': | ||
| os: 'linux' | ||
| arch: 'amd64' | ||
| 'darwin-amd64': | ||
| os: 'darwin' | ||
| arch: 'amd64' | ||
|
|
||
| # included extensions | ||
| extensions: | ||
| - github.com/hpc/kraken/extensions/IPv4 | ||
| - github.com/hpc/kraken/extensions/RPi3 | ||
| - github.com/hpc/kraken/extensions/HostThermal | ||
| - github.com/hpc/kraken/extensions/HostFrequencyScaler | ||
| # included modules | ||
| modules: | ||
| - github.com/hpc/kraken/modules/restapi | ||
| - github.com/hpc/kraken/modules/rfpipower | ||
| - github.com/hpc/kraken/modules/pipxe | ||
| - github.com/hpc/kraken/modules/hostthermaldiscovery | ||
| - github.com/hpc/kraken/modules/hostfrequencyscaling | ||
| - github.com/hpc/kraken/modules/cpuburn | ||
| - github.com/hpc/kraken/modules/websocket | ||
|
|
There was a problem hiding this comment.
This file looks like it's the same as pipxe so it can probably be removed.
Comment on lines
416
to
434
| // // setup a ticker for checking whether PS is enforced in Thermal bound scenario | ||
| // if hfs.cfg.GetThermalBoundScaler() == true { | ||
|
|
||
| // dur, _ := time.ParseDuration("1s") | ||
| // thermalCheckTick := time.NewTicker(dur) | ||
|
|
||
| // // thermal ticker | ||
| // for { | ||
| // select { | ||
| // case <-thermalCheckTick.C: | ||
| // if hfs.psEnforced == true { | ||
| // go hfs.CheckThermalThreshold() | ||
| // } | ||
|
|
||
| // break | ||
| // } | ||
| // } | ||
|
|
||
| // } |
Comment on lines
511
to
532
| // if currentThermal >= thresholdThermal { | ||
| // hfs.mutex.Lock() | ||
| // hfs.psEnforced = true | ||
| // hfs.mutex.Unlock() | ||
| // } else | ||
| if (currentThermal / 1000) < thresholdThermal { | ||
| hfs.mutex.Lock() | ||
| hfs.psEnforced = false | ||
| hfs.mutex.Unlock() | ||
|
|
||
| // url := lib.NodeURLJoin(node.ID().String(), hostFreqScalerURL) | ||
| // ev := core.NewEvent( | ||
| // lib.Event_DISCOVERY, | ||
| // url, | ||
| // &core.DiscoveryEvent{ | ||
| // URL: url, | ||
| // ValueID: profileMap["performance"], | ||
| // }, | ||
| // ) | ||
| // hfs.dchan <- ev | ||
| } | ||
| //hfs.api.Logf(lib.LLERROR, "*** T E M P ***: %v", currentThermal/1000) |
Comment on lines
643
to
654
| // } else { | ||
| // url := lib.NodeURLJoin(node.ID().String(), hostHightoLowFreqScalerURL) | ||
| // ev := core.NewEvent( | ||
| // lib.Event_DISCOVERY, | ||
| // url, | ||
| // &core.DiscoveryEvent{ | ||
| // URL: url, | ||
| // ValueID: currentScalingConfig.CurScalingGovernor, | ||
| // }, | ||
| // ) | ||
| // hfs.dchan <- ev | ||
| // } |
Comment on lines
216
to
219
| // if hostFreqScaler == hostDisc.preFreqScaler { | ||
| // // no change in frequency scaler so no need to generate discovery event | ||
| // return | ||
| // } |
|
|
||
| vid := profileMap[hostFreqScaler] | ||
|
|
||
| //hostDisc.api.Logf(lib., "PRINTSCALER: %s", vid) |
| hostDisc.api.Logf(lib.LLERROR, "Reading CPU thermal sensor failed: %v", err) | ||
| return "" | ||
| } | ||
| //fscalingGovernor := strings.TrimSuffix(string(bscalingGovernor), "\n") |
Comment on lines
353
to
361
| // cpuThermalThresholds := hostDisc.cfg.GetThermalThresholds() | ||
| // lowerNormal := cpuThermalThresholds["CPUThermalThresholds"].GetLowerNormal() | ||
| // upperNormal := cpuThermalThresholds["CPUThermalThresholds"].GetUpperNormal() | ||
|
|
||
| lowerHigh := cpuThermalThresholds["CPUThermalThresholds"].GetLowerHigh() | ||
| upperHigh := cpuThermalThresholds["CPUThermalThresholds"].GetUpperHigh() | ||
| // lowerHigh := cpuThermalThresholds["CPUThermalThresholds"].GetLowerHigh() | ||
| // upperHigh := cpuThermalThresholds["CPUThermalThresholds"].GetUpperHigh() | ||
|
|
||
| lowerCritical := cpuThermalThresholds["CPUThermalThresholds"].GetLowerCritical() | ||
| upperCritical := cpuThermalThresholds["CPUThermalThresholds"].GetUpperCritical() | ||
| // lowerCritical := cpuThermalThresholds["CPUThermalThresholds"].GetLowerCritical() | ||
| // upperCritical := cpuThermalThresholds["CPUThermalThresholds"].GetUpperCritical() |
kpelzel
force-pushed
the
hostfrequencyscaling
branch
from
c07fc7d to
0502246
Compare
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
This module performs mutations related to scaling of CPU frequency to control CPU thermal conditions of HPC node using in-band mechanism.
This module clocks HPC node to "schedutil" scaling governor on boot and whenever CPU temperature reaches to high (warning) condition, module mutates the scaling governor to "powersave". Current implementation handles critical CPU temperature same as high CPU temperature.
Additionally, there are many other mutations intended for different use cases (e.g. switching back to "schedutil" after "powersave") are under considerations and investigations.