- Infos im HLRS Wiki sind nicht rechtsverbindlich und ohne Gewähr -
- Information contained in the HLRS Wiki is not legally binding and HLRS is not responsible for any damages that might result from its use -
User monitoring: Difference between revisions
From HLRS Platforms
Jump to navigationJump to search
| Line 14: | Line 14: | ||
=== How the data is aggregated: === | === How the data is aggregated: === | ||
The total energy consumption per job [in Wh] | |||
# we calculate the median of the power consumption per node over the timeline of the job and sum up the respective contributions from all compute nodes of the job. | # we calculate the median of the power consumption per node over the timeline of the job and sum up the respective contributions from all compute nodes of the job. | ||
# we factor in static contributions from the admin and storage infrastructure, averaged over all compute nodes, assigned per compute node of the job | # we factor in static contributions from the admin and storage infrastructure, averaged over all compute nodes, assigned per compute node of the job | ||
| Line 20: | Line 20: | ||
# we include the overhead corresponding to the efficiency rating of the power supply units | # we include the overhead corresponding to the efficiency rating of the power supply units | ||
The achieved memory bandwidth [in GByte/s]: | |||
# this is the rate at which data can be read from or stored to the main memory. | # this is the rate at which data can be read from or stored to the main memory. | ||
# to calculate the achieved memory bandwidth, we calculate the median of the memory bandwidth per node as reported by Lkiwd [https://github.com/RRZE-HPC/likwid] over the timeline of the job. We then average over the respective contributions from all compute nodes of the job. | # to calculate the achieved memory bandwidth, we calculate the median of the memory bandwidth per node as reported by Lkiwd [https://github.com/RRZE-HPC/likwid] over the timeline of the job. We then average over the respective contributions from all compute nodes of the job. | ||
The achieved floating point performance [in GFlop/s]: | |||
# this is the amount of floating point operations per second. While this metric does not discriminate between single or double precision floating point operations, it will take into account the SIMD width [https://en.wikipedia.org/wiki/Single_instruction,_multiple_data] of the floating point instructions. | # this is the amount of floating point operations per second. While this metric does not discriminate between single or double precision floating point operations, it will take into account the SIMD width [https://en.wikipedia.org/wiki/Single_instruction,_multiple_data] of the floating point instructions. | ||
# to calculate the amount of floating point operations per second, we calculate the median of the amount of floating point operations per second as reported by Lkiwd over the timeline of the job. We then average over the respective contributions from all compute nodes of the job | # to calculate the amount of floating point operations per second, we calculate the median of the amount of floating point operations per second as reported by Lkiwd over the timeline of the job. We then average over the respective contributions from all compute nodes of the job | ||
Revision as of 14:12, 1 August 2024
Long term aggregations for end users
The long term job-specific aggregations for end users consist of eight metrics from the domains of energy consumption, performance characterization and I/O.
- the total energy consumption per job [Wh]
- the achieved memory bandwidth, averaged, per node.
- the achieved floating point performance, averaged, per node.
- the total amount of data written or read to the lustre file systems, per job.
- the achieved peak write bandwidth to the lustre file systems, averaged, per node
- the achieved peak read bandwidth to the lustre file systems, averaged, per node
- the total number of metadata operations on the lustre file systems, per job
- the achieved peak rate of metadata operations on the lustre file systems, averaged, per node
How the data is aggregated:
The total energy consumption per job [in Wh]
- we calculate the median of the power consumption per node over the timeline of the job and sum up the respective contributions from all compute nodes of the job.
- we factor in static contributions from the admin and storage infrastructure, averaged over all compute nodes, assigned per compute node of the job
- we add a static contribution from the cooling distribution units, averaged over all compute nodes, assigned per compute node of the job
- we include the overhead corresponding to the efficiency rating of the power supply units
The achieved memory bandwidth [in GByte/s]:
- this is the rate at which data can be read from or stored to the main memory.
- to calculate the achieved memory bandwidth, we calculate the median of the memory bandwidth per node as reported by Lkiwd [1] over the timeline of the job. We then average over the respective contributions from all compute nodes of the job.
The achieved floating point performance [in GFlop/s]:
- this is the amount of floating point operations per second. While this metric does not discriminate between single or double precision floating point operations, it will take into account the SIMD width [2] of the floating point instructions.
- to calculate the amount of floating point operations per second, we calculate the median of the amount of floating point operations per second as reported by Lkiwd over the timeline of the job. We then average over the respective contributions from all compute nodes of the job