spdf/elektron
Archived
1
0
Fork 0
Code Issues 4 Pull requests 1 Projects Releases Packages Wiki Activity

Comments: Explain what Variance is used for.

Browse source
This commit is contained in:
Akash Kothawale 2018-02-05 16:01:48 -05:00 • committed by Pradyumna Kaushik
parent 2d3158d99a
commit acf9332a50
1 changed files with 8 additions and 0 deletions
Download patch file Download diff file Expand all files Collapse all files

8
pcp/pcp.go
View file

@ -62,7 +62,9 @@ func Start(quit chan struct{}, logging *bool, logMType chan elecLogDef.LogMessag
for i := 0; i < 8; i++ {
host := fmt.Sprintf("stratos-00%d.cs.binghamton.edu", i+1)
if slaveID, ok := baseSchedRef.HostNameToSlaveID[host]; ok {
baseSchedRef.TasksRunningMutex.Lock()
tasksRunning := len(baseSchedRef.Running[slaveID])
baseSchedRef.TasksRunningMutex.Unlock()
if tasksRunning > 0 {
cpuTaskShares[i] = cpuUtils[i] / float64(tasksRunning)
memTaskShares[i] = memUtils[i] / float64(tasksRunning)
@ -70,6 +72,12 @@ func Start(quit chan struct{}, logging *bool, logMType chan elecLogDef.LogMessag
}
}
// Variance in resource utilization shows how the current workload has been distributed.
// However, if the number of tasks running are not equally distributed, utilization variance figures become
// less relevant as they do not express the distribution of CPU intensive tasks.
// We thus also calculate `task share variance`, which basically signifies how the workload is distributed
// across each node per share.
cpuVariance, _ := stats.Variance(cpuUtils)
cpuTaskSharesVariance, _ := stats.Variance(cpuTaskShares)
memVariance, _ := stats.Variance(memUtils)