Files
nomad/client/stats/cpu.go
Seth Hoenig fd900d0723 client/fingerprint: correctly fingerprint E/P cores of Apple Silicon chips (#16672)
* client/fingerprint: correctly fingerprint E/P cores of Apple Silicon chips

This PR adds detection of asymetric core types (Power & Efficiency) (P/E)
when running on M1/M2 Apple Silicon CPUs. This functionality is provided
by shoenig/go-m1cpu which makes use of the Apple IOKit framework to read
undocumented registers containing CPU performance data. Currently working
on getting that functionality merged upstream into gopsutil, but gopsutil
would still not support detecting P vs E cores like this PR does.

Also refactors the CPUFingerprinter code to handle the mixed core
types, now setting power vs efficiency cpu attributes.

For now the scheduler is still unaware of mixed core types - on Apple
platforms tasks cannot reserve cores anyway so it doesn't matter, but
at least now the total CPU shares available will be correct.

Future work should include adding support for detecting P/E cores on
the latest and upcoming Intel chips, where computation of total cpu shares
is currently incorrect. For that, we should also include updating the
scheduler to be core-type aware, so that tasks of resources.cores on Linux
platforms can be assigned the correct number of CPU shares for the core
type(s) they have been assigned.

node attributes before

cpu.arch                  = arm64
cpu.modelname             = Apple M2 Pro
cpu.numcores              = 12
cpu.reservablecores       = 0
cpu.totalcompute          = 1000

node attributes after

cpu.arch                  = arm64
cpu.frequency.efficiency  = 2424
cpu.frequency.power       = 3504
cpu.modelname             = Apple M2 Pro
cpu.numcores.efficiency   = 4
cpu.numcores.power        = 8
cpu.reservablecores       = 0
cpu.totalcompute          = 37728

* fingerprint/cpu: follow up cr items
2023-03-28 08:27:58 -05:00

91 lines
2.2 KiB
Go

package stats
import (
"runtime"
"time"
shelpers "github.com/hashicorp/nomad/helper/stats"
"github.com/shirou/gopsutil/v3/cpu"
)
// CpuStats calculates cpu usage percentage
type CpuStats struct {
prevCpuTime float64
prevTime time.Time
totalCpus int
}
// NewCpuStats returns a cpu stats calculator
func NewCpuStats() *CpuStats {
numCpus := runtime.NumCPU()
cpuStats := &CpuStats{
totalCpus: numCpus,
}
return cpuStats
}
// Percent calculates the cpu usage percentage based on the current cpu usage
// and the previous cpu usage where usage is given as time in nanoseconds spend
// in the cpu
func (c *CpuStats) Percent(cpuTime float64) float64 {
now := time.Now()
if c.prevCpuTime == 0.0 {
// invoked first time
c.prevCpuTime = cpuTime
c.prevTime = now
return 0.0
}
timeDelta := now.Sub(c.prevTime).Nanoseconds()
ret := c.calculatePercent(c.prevCpuTime, cpuTime, timeDelta)
c.prevCpuTime = cpuTime
c.prevTime = now
return ret
}
// TicksConsumed calculates the total ticks consumes by the process across all
// cpu cores
func (c *CpuStats) TicksConsumed(percent float64) float64 {
return (percent / 100) * float64(shelpers.TotalTicksAvailable()) / float64(c.totalCpus)
}
func (c *CpuStats) calculatePercent(t1, t2 float64, timeDelta int64) float64 {
vDelta := t2 - t1
if timeDelta <= 0 || vDelta <= 0.0 {
return 0.0
}
overall_percent := (vDelta / float64(timeDelta)) * 100.0
return overall_percent
}
func (h *HostStatsCollector) collectCPUStats() (cpus []*CPUStats, totalTicks float64, err error) {
ticksConsumed := 0.0
cpuStats, err := cpu.Times(true)
if err != nil {
return nil, 0.0, err
}
cs := make([]*CPUStats, len(cpuStats))
for idx, cpuStat := range cpuStats {
percentCalculator, ok := h.statsCalculator[cpuStat.CPU]
if !ok {
percentCalculator = NewHostCpuStatsCalculator()
h.statsCalculator[cpuStat.CPU] = percentCalculator
}
idle, user, system, total := percentCalculator.Calculate(cpuStat)
cs[idx] = &CPUStats{
CPU: cpuStat.CPU,
User: user,
System: system,
Idle: idle,
Total: total,
}
ticksConsumed += (total / 100.0) * (float64(shelpers.TotalTicksAvailable()) / float64(len(cpuStats)))
}
return cs, ticksConsumed, nil
}