Event query discovery from Google cluster traces · run 2026-10-02 22:50–2026-10-03 17:01 CEST Workflow Run uri icon

start time

  • October 2, 2026 10:50 PM

end time

  • October 3, 2026 5:01 PM

duration (s)

  • 65456

duration time calculation method

  • Slurm ElapsedRaw of the job (start to end). Display timestamps use Europe/Berlin.

run status

  • Succeeded

CPU efficiency (%)

  • 18.1

data read (GB)

  • 144.74336534528

data written (GB)

  • 2.85314383872

CPU time (s)

  • 71153.0

CPU time calculation method

  • Slurm TotalCPU of the job (user + system CPU time of all steps); 6 CPUs allocated of 256 node CPUs.

task count

  • 3

peak memory (GB)

  • 0.542294016

GPU usage status

  • CPU-only job

memory calculation method

  • Peak memory: the largest resident memory (Slurm MaxRSS) of any job step, converted from bytes to gigabytes. No average memory: Slurm does not record memory use over time.

energy consumed (kWh)

  • 0.1488196799550822

energy measurement coverage

  • 1.5% of the node (shared, estimate)

node power average (W)

  • 578.1

node power maximum (W)

  • 726.0

node sleeping idle power (W)

  • 393.0

carbon emission (kg CO2e)

  • 0.06828332241728056

carbon intensity (kg CO2e/kWh)

  • 0.45883261164041145

energy metric source

  • Slurm acct_gather_energy/ipmi; Prometheus node_exporter and ipmi_exporter on the compute node

energy calculation method

  • Node energy: Slurm acct_gather_energy/ipmi ConsumedEnergyRaw, 35942944 J over the job (65456 s, whole node). Job share = job CPU time 71153.0 s / CPU time of all processes on the node in the same interval 4773577.7 s (node_exporter node_cpu_seconds_total, all modes except idle and iowait; 6547 samples about every 10 s, covering 100% of the job) = 0.0149. Job energy = 35942944 J x 0.0149 = 535751 J, converted to kilowatt-hours (1 kWh = 3,600,000 J): 0.148820 kWh. Single node hpc-cms01-002 (HPC@HU), shared with other jobs. Energy is an estimate: the job's CPU-time share of the measured whole-node energy. Idle and other node power is distributed by CPU time; cooling and network are excluded.

carbon calculation method

  • Time-matched: the run's energy 0.148820 kWh x the 15-minute carbon intensity of the grid (Electricity Maps, zone DE, lifecycle, flow-traced) in the quarter-hours the run used, weighted by the run's energy in each quarter-hour (split by power readings; 74 quarter-hours from 2026-10-02T20:45 to 2026-10-03T15:15 UTC, 216 to 599 g/kWh) = 458.8 g/kWh, so 68.28 g CO2e. Window start and end are Europe/Berlin times. Recalculated on 2026-11-07: replaces the earlier value 68.26 g CO2e (458.7 g/kWh, Electricity Maps).

idle power method

  • Node hpc-cms01-002 saves energy when nothing runs: its idle processor cores sleep (low-power states), while the node itself stays on and takes new jobs at once. Sleeping idle power: 393 W, the median IPMI power of the node in the cluster monitoring (Prometheus, a reading every 5 s) over the 30 days up to 2026-10-30, counting only times when nothing had run on the node for at least 30 minutes (load below 0.2, both GPUs below 60 W). No energy above idle is given for this run: the node's cores sleep when idle, and its idle power without sleep is not known. The node is assumed to have had its current power profile also on the day of the run.

git commit

  • c898ae4fda50a3463e992344c946f9b983fae7e7

job name

  • event-query-discovery (Slurm job 1598478, partition standard, exit code 0:0)

code version

  • sha256:89d9538633ba919f433b818e2dc32398966c4b5e8f71cef7553dde493c1a6625

trace types

  • Slurm accounting (sacct); node CPU and IPMI power samples; node hardware inventory; stage records (times, exit status, CPU time, peak memory)

trace data format

  • TSV, JSON and plain text

node CPU utilisation average (%)

  • 28.5

node conditions method

  • Node CPU utilisation: CPU time of all processes on the node (node_exporter, all modes except idle and iowait) / (256 CPUs x sampled time) = 28.5% (72.9 of 256 CPUs busy on average). Node power: IPMI power of the node (ipmi_exporter), 6547 readings about every 10 s: average 578 W (over time), lowest 486 W, highest 726 W. CPU efficiency: job CPU time 71153 s / (6 allocated CPUs x 65456 s) = 18.1%. Data read and written: Slurm accounting MaxDiskRead and MaxDiskWrite of the job's 1 step(s): 144.743 GB read, 2.853 GB written. This is all data the job's processes read and wrote, on any file system.

workflow engine

execution host

  • hpc-cms01-002

node allocatable CPU

  • 256

node allocatable memory (GB)

  • 1081.494478848

node architecture

  • x86_64; 2 x AMD EPYC 7713 64-Core Processor; 256 hardware threads

os image

  • openSUSE Leap 16.0

kernel version

  • 6.12.0-160000.36-default

backend