ClusterCockpit/cc-metric-collector
1
0
Fork 0
mirror of https://github.com/ClusterCockpit/cc-metric-collector.git synced 2025-05-11 20:31:43 +02:00
Code Issues Packages Projects Releases Wiki Activity

Add energy metrics from NVML to Nvidia NVML collector

Browse Source
This commit is contained in:
Thomas Roehl 2025-04-28 15:29:13 +00:00
parent f8b2ac0d2c
commit a606a3af01
1 changed files with 86 additions and 27 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

113
collectors/nvidiaMetric.go
View File

@ -27,10 +27,12 @@ type NvidiaCollectorConfig struct {
} }
type NvidiaCollectorDevice struct { type NvidiaCollectorDevice struct {
device nvml.Device device nvml.Device
excludeMetrics map[string]bool excludeMetrics map[string]bool
tags map[string]string tags map[string]string
meta map[string]string meta map[string]string
lastEnergyReading uint64
lastEnergyTimestamp time.Time
} }
type NvidiaCollector struct { type NvidiaCollector struct {
@ -149,6 +151,8 @@ func (m *NvidiaCollector) Init(config json.RawMessage) error {
// Add device handle // Add device handle
g.device = device g.device = device
g.lastEnergyReading = 0
g.lastEnergyTimestamp = time.Now()
// Add tags // Add tags
g.tags = map[string]string{ g.tags = map[string]string{
@ -206,7 +210,7 @@ func (m *NvidiaCollector) Init(config json.RawMessage) error {
return nil return nil
} }
func readMemoryInfo(device NvidiaCollectorDevice, output chan lp.CCMessage) error { func readMemoryInfo(device *NvidiaCollectorDevice, output chan lp.CCMessage) error {
if !device.excludeMetrics["nv_fb_mem_total"] || !device.excludeMetrics["nv_fb_mem_used"] || !device.excludeMetrics["nv_fb_mem_reserved"] { if !device.excludeMetrics["nv_fb_mem_total"] || !device.excludeMetrics["nv_fb_mem_used"] || !device.excludeMetrics["nv_fb_mem_reserved"] {
var total uint64 var total uint64
var used uint64 var used uint64
@ -250,7 +254,7 @@ func readMemoryInfo(device NvidiaCollectorDevice, output chan lp.CCMessage) erro
return nil return nil
} }
func readBarMemoryInfo(device NvidiaCollectorDevice, output chan lp.CCMessage) error { func readBarMemoryInfo(device *NvidiaCollectorDevice, output chan lp.CCMessage) error {
if !device.excludeMetrics["nv_bar1_mem_total"] || !device.excludeMetrics["nv_bar1_mem_used"] { if !device.excludeMetrics["nv_bar1_mem_total"] || !device.excludeMetrics["nv_bar1_mem_used"] {
meminfo, ret := nvml.DeviceGetBAR1MemoryInfo(device.device) meminfo, ret := nvml.DeviceGetBAR1MemoryInfo(device.device)
if ret != nvml.SUCCESS { if ret != nvml.SUCCESS {
@ -277,7 +281,7 @@ func readBarMemoryInfo(device NvidiaCollectorDevice, output chan lp.CCMessage) e
return nil return nil
} }
func readUtilization(device NvidiaCollectorDevice, output chan lp.CCMessage) error { func readUtilization(device *NvidiaCollectorDevice, output chan lp.CCMessage) error {
isMig, ret := nvml.DeviceIsMigDeviceHandle(device.device) isMig, ret := nvml.DeviceIsMigDeviceHandle(device.device)
if ret != nvml.SUCCESS { if ret != nvml.SUCCESS {
err := errors.New(nvml.ErrorString(ret)) err := errors.New(nvml.ErrorString(ret))
@ -319,7 +323,7 @@ func readUtilization(device NvidiaCollectorDevice, output chan lp.CCMessage) err
return nil return nil
} }
func readTemp(device NvidiaCollectorDevice, output chan lp.CCMessage) error { func readTemp(device *NvidiaCollectorDevice, output chan lp.CCMessage) error {
if !device.excludeMetrics["nv_temp"] { if !device.excludeMetrics["nv_temp"] {
// Retrieves the current temperature readings for the device, in degrees C. // Retrieves the current temperature readings for the device, in degrees C.
// //
@ -338,7 +342,7 @@ func readTemp(device NvidiaCollectorDevice, output chan lp.CCMessage) error {
return nil return nil
} }
func readFan(device NvidiaCollectorDevice, output chan lp.CCMessage) error { func readFan(device *NvidiaCollectorDevice, output chan lp.CCMessage) error {
if !device.excludeMetrics["nv_fan"] { if !device.excludeMetrics["nv_fan"] {
// Retrieves the intended operating speed of the device's fan. // Retrieves the intended operating speed of the device's fan.
// //
@ -361,7 +365,7 @@ func readFan(device NvidiaCollectorDevice, output chan lp.CCMessage) error {
return nil return nil
} }
// func readFans(device NvidiaCollectorDevice, output chan lp.CCMessage) error { // func readFans(device *NvidiaCollectorDevice, output chan lp.CCMessage) error {
// if !device.excludeMetrics["nv_fan"] { // if !device.excludeMetrics["nv_fan"] {
// numFans, ret := nvml.DeviceGetNumFans(device.device) // numFans, ret := nvml.DeviceGetNumFans(device.device)
// if ret == nvml.SUCCESS { // if ret == nvml.SUCCESS {
@ -382,7 +386,7 @@ func readFan(device NvidiaCollectorDevice, output chan lp.CCMessage) error {
// return nil // return nil
// } // }
func readEccMode(device NvidiaCollectorDevice, output chan lp.CCMessage) error { func readEccMode(device *NvidiaCollectorDevice, output chan lp.CCMessage) error {
if !device.excludeMetrics["nv_ecc_mode"] { if !device.excludeMetrics["nv_ecc_mode"] {
// Retrieves the current and pending ECC modes for the device. // Retrieves the current and pending ECC modes for the device.
// //
@ -416,7 +420,7 @@ func readEccMode(device NvidiaCollectorDevice, output chan lp.CCMessage) error {
return nil return nil
} }
func readPerfState(device NvidiaCollectorDevice, output chan lp.CCMessage) error { func readPerfState(device *NvidiaCollectorDevice, output chan lp.CCMessage) error {
if !device.excludeMetrics["nv_perf_state"] { if !device.excludeMetrics["nv_perf_state"] {
// Retrieves the current performance state for the device. // Retrieves the current performance state for the device.
// //
@ -436,13 +440,16 @@ func readPerfState(device NvidiaCollectorDevice, output chan lp.CCMessage) error
return nil return nil
} }
func readPowerUsage(device NvidiaCollectorDevice, output chan lp.CCMessage) error { func readPowerUsage(device *NvidiaCollectorDevice, output chan lp.CCMessage) error {
if !device.excludeMetrics["nv_power_usage"] { if !device.excludeMetrics["nv_power_usage"] {
// Retrieves power usage for this GPU in milliwatts and its associated circuitry (e.g. memory) // Retrieves power usage for this GPU in milliwatts and its associated circuitry (e.g. memory)
// //
// On Fermi and Kepler GPUs the reading is accurate to within +/- 5% of current power draw. // On Fermi and Kepler GPUs the reading is accurate to within +/- 5% of current power draw.
// On Ampere (except GA100) or newer GPUs, the API returns power averaged over 1 sec interval.
// On GA100 and older architectures, instantaneous power is returned.
// //
// It is only available if power management mode is supported // It is only available if power management mode is supported.
mode, ret := nvml.DeviceGetPowerManagementMode(device.device) mode, ret := nvml.DeviceGetPowerManagementMode(device.device)
if ret != nvml.SUCCESS { if ret != nvml.SUCCESS {
return nil return nil
@ -461,7 +468,54 @@ func readPowerUsage(device NvidiaCollectorDevice, output chan lp.CCMessage) erro
return nil return nil
} }
func readClocks(device NvidiaCollectorDevice, output chan lp.CCMessage) error { func readEnergyConsumption(device *NvidiaCollectorDevice, output chan lp.CCMessage) error {
// Retrieves total energy consumption for this GPU in millijoules (mJ) since the driver was last reloaded
// For Volta or newer fully supported devices.
if (!device.excludeMetrics["nv_energy"]) && (!device.excludeMetrics["nv_energy_abs"]) && (!device.excludeMetrics["nv_average_power"]) {
now := time.Now()
mode, ret := nvml.DeviceGetPowerManagementMode(device.device)
if ret != nvml.SUCCESS {
return nil
}
if mode == nvml.FEATURE_ENABLED {
energy, ret := nvml.DeviceGetTotalEnergyConsumption(device.device)
if ret == nvml.SUCCESS {
if device.lastEnergyReading != 0 {
if !device.excludeMetrics["nv_energy"] {
y, err := lp.NewMetric("nv_energy", device.tags, device.meta, (energy-device.lastEnergyReading)/1000, now)
if err == nil {
y.AddMeta("unit", "Joules")
output <- y
}
}
if !device.excludeMetrics["nv_average_power"] {
energyDiff := (energy - device.lastEnergyReading) / 1000
timeDiff := now.Sub(device.lastEnergyTimestamp)
y, err := lp.NewMetric("nv_average_power", device.tags, device.meta, energyDiff/uint64(timeDiff.Seconds()), now)
if err == nil {
y.AddMeta("unit", "watts")
output <- y
}
}
}
if !device.excludeMetrics["nv_energy_abs"] {
y, err := lp.NewMetric("nv_energy_abs", device.tags, device.meta, energy/1000, now)
if err == nil {
y.AddMeta("unit", "Joules")
output <- y
}
}
device.lastEnergyReading = energy
device.lastEnergyTimestamp = time.Now()
}
}
}
return nil
}
func readClocks(device *NvidiaCollectorDevice, output chan lp.CCMessage) error {
// Retrieves the current clock speeds for the device. // Retrieves the current clock speeds for the device.
// //
// Available clock information: // Available clock information:
@ -513,7 +567,7 @@ func readClocks(device NvidiaCollectorDevice, output chan lp.CCMessage) error {
return nil return nil
} }
func readMaxClocks(device NvidiaCollectorDevice, output chan lp.CCMessage) error { func readMaxClocks(device *NvidiaCollectorDevice, output chan lp.CCMessage) error {
// Retrieves the maximum clock speeds for the device. // Retrieves the maximum clock speeds for the device.
// //
// Available clock information: // Available clock information:
@ -571,7 +625,7 @@ func readMaxClocks(device NvidiaCollectorDevice, output chan lp.CCMessage) error
return nil return nil
} }
func readEccErrors(device NvidiaCollectorDevice, output chan lp.CCMessage) error { func readEccErrors(device *NvidiaCollectorDevice, output chan lp.CCMessage) error {
if !device.excludeMetrics["nv_ecc_uncorrected_error"] { if !device.excludeMetrics["nv_ecc_uncorrected_error"] {
// Retrieves the total ECC error counts for the device. // Retrieves the total ECC error counts for the device.
// //
@ -602,7 +656,7 @@ func readEccErrors(device NvidiaCollectorDevice, output chan lp.CCMessage) error
return nil return nil
} }
func readPowerLimit(device NvidiaCollectorDevice, output chan lp.CCMessage) error { func readPowerLimit(device *NvidiaCollectorDevice, output chan lp.CCMessage) error {
if !device.excludeMetrics["nv_power_max_limit"] { if !device.excludeMetrics["nv_power_max_limit"] {
// Retrieves the power management limit associated with this device. // Retrieves the power management limit associated with this device.
// //
@ -622,7 +676,7 @@ func readPowerLimit(device NvidiaCollectorDevice, output chan lp.CCMessage) erro
return nil return nil
} }
func readEncUtilization(device NvidiaCollectorDevice, output chan lp.CCMessage) error { func readEncUtilization(device *NvidiaCollectorDevice, output chan lp.CCMessage) error {
isMig, ret := nvml.DeviceIsMigDeviceHandle(device.device) isMig, ret := nvml.DeviceIsMigDeviceHandle(device.device)
if ret != nvml.SUCCESS { if ret != nvml.SUCCESS {
err := errors.New(nvml.ErrorString(ret)) err := errors.New(nvml.ErrorString(ret))
@ -649,7 +703,7 @@ func readEncUtilization(device NvidiaCollectorDevice, output chan lp.CCMessage)
return nil return nil
} }
func readDecUtilization(device NvidiaCollectorDevice, output chan lp.CCMessage) error { func readDecUtilization(device *NvidiaCollectorDevice, output chan lp.CCMessage) error {
isMig, ret := nvml.DeviceIsMigDeviceHandle(device.device) isMig, ret := nvml.DeviceIsMigDeviceHandle(device.device)
if ret != nvml.SUCCESS { if ret != nvml.SUCCESS {
err := errors.New(nvml.ErrorString(ret)) err := errors.New(nvml.ErrorString(ret))
@ -676,7 +730,7 @@ func readDecUtilization(device NvidiaCollectorDevice, output chan lp.CCMessage)
return nil return nil
} }
func readRemappedRows(device NvidiaCollectorDevice, output chan lp.CCMessage) error { func readRemappedRows(device *NvidiaCollectorDevice, output chan lp.CCMessage) error {
if !device.excludeMetrics["nv_remapped_rows_corrected"] || if !device.excludeMetrics["nv_remapped_rows_corrected"] ||
!device.excludeMetrics["nv_remapped_rows_uncorrected"] || !device.excludeMetrics["nv_remapped_rows_uncorrected"] ||
!device.excludeMetrics["nv_remapped_rows_pending"] || !device.excludeMetrics["nv_remapped_rows_pending"] ||
@ -729,7 +783,7 @@ func readRemappedRows(device NvidiaCollectorDevice, output chan lp.CCMessage) er
return nil return nil
} }
func readProcessCounts(device NvidiaCollectorDevice, output chan lp.CCMessage) error { func readProcessCounts(device *NvidiaCollectorDevice, output chan lp.CCMessage) error {
if !device.excludeMetrics["nv_compute_processes"] { if !device.excludeMetrics["nv_compute_processes"] {
// Get information about processes with a compute context on a device // Get information about processes with a compute context on a device
// //
@ -821,7 +875,7 @@ func readProcessCounts(device NvidiaCollectorDevice, output chan lp.CCMessage) e
return nil return nil
} }
func readViolationStats(device NvidiaCollectorDevice, output chan lp.CCMessage) error { func readViolationStats(device *NvidiaCollectorDevice, output chan lp.CCMessage) error {
var violTime nvml.ViolationTime var violTime nvml.ViolationTime
var ret nvml.Return var ret nvml.Return
@ -935,7 +989,7 @@ func readViolationStats(device NvidiaCollectorDevice, output chan lp.CCMessage)
return nil return nil
} }
func readNVLinkStats(device NvidiaCollectorDevice, output chan lp.CCMessage) error { func readNVLinkStats(device *NvidiaCollectorDevice, output chan lp.CCMessage) error {
// Retrieves the specified error counter value // Retrieves the specified error counter value
// Please refer to \a nvmlNvLinkErrorCounter_t for error counters that are available // Please refer to \a nvmlNvLinkErrorCounter_t for error counters that are available
// //
@ -1070,7 +1124,7 @@ func (m *NvidiaCollector) Read(interval time.Duration, output chan lp.CCMessage)
return return
} }
readAll := func(device NvidiaCollectorDevice, output chan lp.CCMessage) { readAll := func(device *NvidiaCollectorDevice, output chan lp.CCMessage) {
name, ret := nvml.DeviceGetName(device.device) name, ret := nvml.DeviceGetName(device.device)
if ret != nvml.SUCCESS { if ret != nvml.SUCCESS {
name = "NoName" name = "NoName"
@ -1110,6 +1164,11 @@ func (m *NvidiaCollector) Read(interval time.Duration, output chan lp.CCMessage)
cclog.ComponentDebug(m.name, "readPowerUsage for device", name, "failed") cclog.ComponentDebug(m.name, "readPowerUsage for device", name, "failed")
} }
err = readEnergyConsumption(device, output)
if err != nil {
cclog.ComponentDebug(m.name, "readEnergyConsumption for device", name, "failed")
}
err = readClocks(device, output) err = readClocks(device, output)
if err != nil { if err != nil {
cclog.ComponentDebug(m.name, "readClocks for device", name, "failed") cclog.ComponentDebug(m.name, "readClocks for device", name, "failed")
@ -1169,7 +1228,7 @@ func (m *NvidiaCollector) Read(interval time.Duration, output chan lp.CCMessage)
// Actual read loop over all attached Nvidia GPUs // Actual read loop over all attached Nvidia GPUs
for i := 0; i < m.num_gpus; i++ { for i := 0; i < m.num_gpus; i++ {
readAll(m.gpus[i], output) readAll(&m.gpus[i], output)
// Iterate over all MIG devices if any // Iterate over all MIG devices if any
if m.config.ProcessMigDevices { if m.config.ProcessMigDevices {
@ -1243,7 +1302,7 @@ func (m *NvidiaCollector) Read(interval time.Duration, output chan lp.CCMessage)
} }
} }
readAll(migDevice, output) readAll(&migDevice, output)
} }
} }
} }